TFA - Cohen PDF
TFA - Cohen PDF
TFA - Cohen PDF
ANALYSIS
Leon Cohen
Hunter College and Graduate Center of
The City University of New York
The publisher offers discounts on this book when ordered in hulk quantities. For
more information, contact: Corporate Sales Department. Prentice Hall PTR.
I Lake Street, Upper Saddle River, New Jersey, 07694. Phone 800-382-3419.
Fax 201-236-7141 E-mail: [email protected]
10 987654
ISBN 0-13-594532-1
.........................................................
1.2 Time Description of Signals ........................................... 2
1.3 Frequency Description of Signals ..................................... 6
1.4 Simple Calculation Tricks ............................................. 8
1.5 Bandwidth Equation ................................................. 15
vii
Viii Contents
18. Scale
18.1 Introduction ..................................... 242
18.2 The Scale and Compression Operator ............. 242
18.3 The Scale Eigenfunctions ......................... 244
18.4 The Scale Transform .............................. 245
18.5 Signals with High Scale Content .................. 248
18.6 Scale Characteristic Function ..................... 249
18.7 Mean Scale and Bandwidth ....................... 250
18.8 Instantaneous Scale .............................. 251
18.9 Uncertainty Principle for Scale .................... 251
18.10 Frequency and Other Scaling ..................... 252
18.11 Appendix ........................................ 253
Leon Cohen
New York
xiii
Notation in Brief
The main notational conventions we use are as follows:
1. Integrals. All integrals without limits imply integration from minus -oo to oc,
f t
2. Fourier Transform Pairs. We use s(t) to denote a signal and S(w) its Fourier
transform and normalize symmetrically:
; s(t) =
S(w) 2=r 1 s(t) e-2-tdt = f S(w) e'-tdw
vf2- 7r
For other quantities there will be exceptions in regard to the factor of 2n due to
historically accepted conventions. We consistently use angular frequency, w.
3. Magnitude and Phase. It is often advantageous to express a signal and its Fourier
transform in terms of their respective amplitudes and phases. The notation we use
is
s(t) = A(t)e.i'v(t) ; S(w) = B(w)do(w)
and we use "phase" and "spectral phase" to denote cp and O(w), respectively, and
"amplitude" and "spectral amplitude" to denote A(t) and B(w).
4. Functions. We often use the variable to denote a function. That is, f (x) and f (y)
may not necessarily be the same function, the individuality of the functions being
denoted by the variable, x or y. Where confusion may arise we will use different
notation to emphasize the distinction.
5. Averages. Global and conditional averages are denoted by the following con-
ventions:
(w) e.g., average weight
(w) h e.g., average weight for a given height
0-2. - (w)2
(w2) e.g., standard deviation of weight
Qwlh=(w2)h-(w)h e.g., standard deviation of weight for a given height
6. Operators. Symbols in calligraphic letters are operators. For example, the fre-
quency operator, W, and time operator, T, are
T=j d
W=-3dt &AJ
xv
Time-Frequency Analysis
Chapter 1
1.1 INTRODUCTION
In this chapter we develop the basic ideas of time and frequency analysis from the
perspective that the standard treatments already contain the seeds and motivations
for the necessity of a combined time-frequency description. Signal analysis is the
study and characterization of the basic properties of signals and was historically
developed concurrently with the discovery of the fundamental signals in nature,
such as the electric field, sound wave, and electric currents. A signal is generally a
function of many variables. For example, the electric field varies in both space and
time. Our main emphasis will be the time variation, although the ideas developed
are easily extended to spatial and other variables, and we do so in the latter part of
the book
The time variation of a signal is fundamental because time is fundamental. How-
ever, if we want to gain more understanding, it is often advantageous to study
the signal in a different representation. This is done by expanding the signal in
a complete set of functions, and from a mathematical point of view there are an
infinite number of ways this can be done. What makes a particular representation
important is that the characteristics of the signal are understood better in that rep-
resentation because the representation is characterized by a physical quantity that
is important in nature or for the situation at hand. Besides time, the most important
representation is frequency. The mathematics of the frequency representation was
invented by Fourier, whose main motivation was to find the equation governing
the behavior of heat. The contributions of Fourier were milestones because indeed
he did find the fundamental equation governing heat, and, in addition, he invented
the remarkable mathematics to handle discontinuities (1807). He had to be able to
I
2 Chap. 1 The Time and Frequency Description of Signals
handle discontinuities because in one of the most basic problems regarding heat,
namely, when hot and cold objects are put in contact, a discontinuity in temper-
ature arises. Founer's idea, that a discontinuous function can be expressed as the
sum of continuous functions - an absurd idea on the face of it, which the great
scientists of that time, including Laplace and Lagrange, did not hesitate to call ab-
surd in nondiplomatic language - turned out to be one of the great innovations of
mathematics and science.' However, the reason spectral analysis is one of the most
powerful scientific methods ever discovered is due to the contributions of Bunsen
and Kirchhoff about sixty years after Fourier presented his ideas (1807) and about 35
years after his death in 1830. Spectral analysis turned out to be much more impor-
tant than anyone in Fourier's time could have envisioned. This came about with
the invention of the spectroscope2 and with the discovery that by spectrally ana-
lyzing light we can determine the nature of matter; that atoms and molecules are
fingerprinted by the frequency spectrum of the light they emit. This is the modern
usage of spectral analysis. Its discoverers, Bunsen and Kirchhoff, observed (around
1865) that light spectra can be used for recognition, detection, and classification of
substances because they are unique to each substance.
This idea, along with its extension to other waveforms and the invention of the
tools needed to carry out spectral decomposition, certainly ranks as one of the most
important discoveries in the history of mankind. It could certainly be argued that
the spectroscope and its variations are the most important scientific tools ever de-
vised. The analysis of spectra has led to the discovery of the basic laws of nature and
has allowed us to understand the composition and nature of substances on earth
and in stars millions of light years away. It would be appropriate to refer to spectral
analysis as Bunsen-Kirchhoff analysis.
We say that such a signal is of constant amplitude. This does not mean that the
signal is of constant value, but that the maxima and minima of the oscillations are
constant. The frequency, wo, has a clear physical interpretation, namely the number
of oscillations, or ups and downs, per unit time.
One attempts to generalize the simplicity of the sinusoid by hoping that a gen-
eral signal can be written in the form
s(t) = a(t) cost9(t) (1.2)
where the amplitude, a(t), and phase, i9(t), are now arbitrary functions of time. To
emphasize that they generally change in time, the phrases amplitude modulation
and phase modulation are often used, since the word modulation means change.
Difficulties arise immediately. Nature does not break up a signal for us in terms
of amplitude and phase. Nature only gives us the left-hand side, s(t). Even if the
signal were generated by a human by way of Eq. (1.2) with specific amplitude and
phase functions, that particular a(t) and t9(t) would not be special since there are
an infinite number of ways of choosing different pairs of amplitudes and phases
that generate the same signal. Is there one pair that is special?
Also, it is often advantageous to write a signal in complex form
and we want to take the actual signal at hand to be the real part of the complex sig-
nal. How do we choose A and cp or, equivalently, how do we choose the imaginary
part, s;? It is important to realize that the phase and amplitude of the real signal
are not generally the same as the phase and amplitude of the complex signal. We
have emphasized this by using different symbols for the phases and amplitudes in
Eqs. (1.2) and (1.3).
How to unambiguously define amplitude and phase and how to define a com-
plex signal corresponding to a real signal will be the subject of the next chapter.
From the ideas and mathematics developed in this chapter we will see why defin-
ing a complex signal is advantageous and we will lay the groundwork to see how
to do it. In this chapter we consider complex signals but make no assumptions re-
garding the amplitude and phase.
Energy Density or Instantaneous Power. How much energy a signal has and specif-
ically how much energy it takes to produce it is a central idea. In the case of elec-
tromagnetic theory, the electric energy density is the absolute square of the elec-
tric field and similarly for the magnetic field. This was derived by Poynting using
Maxwell's equations and is known as Poynting's theorem. In circuits, the energy
4 Chap. 1 The Time and Frequency Description of Signals
density is proportional to the voltage squared. For a sound wave it is the pressure
squared. Therefore, the energy or intensity of a signal is generally I s(t) 12. That is,
in a small interval of time, At, it takes I s(t) J2ot amount of energy to produce the
signal at that time. Since I s(t) I2 is the energy per unit time it may be appropriately
called the energy density or the instantaneous power since power is the amount of
work per unit time. Therefore
12
1 s(t) = energy or intensity per unit time at time t
(energy density or instantaneous power)
Signal analysis has been extended to many diverse types of data, including eco-
nomical and sociological. It is certainly not obvious that in those cases we can mean-
ingfully talk about the energy density per unit time and take I s(t) I2 to be its value.
However, that is what is done by "analogy", which is appropriate if the results are
fruitful.
Total Energy. If I s(t) I 2 is the energy per unit time, then the total energy is obtained
by summing or integrating over all time,
E _ f I s(t) I2
dt (1.4)
For signals with finite energy we can take, without loss of generality, the total en-
ergy to be equal to one. For many signals the total energy is infinite. For example, a
pure sine wave has infinite total energy, which is reasonable since to keep on pro-
ducing it, work must be expended continually. Such cases can usually be handled
without much difficulty by a limiting process.
(t) = f (1.5)
The reasons for defining an average are that it may give a gross characterization
of the density and it may give an indication of where the density is concentrated.
Many measures can be used to ascertain whether the density is concentrated around
the average, the most common being the standard deviation, art, given by
Note that for a complex signal, time averages depend only on the amplitude.
(t) _ V 7r
-.(t-t0)2
dt = to (1.10)
a 2-
(t) = V;- / t e "
2 2
+ to (1.11)
( o) dt = 2a
Hence
a2 = (t2)-(t)2=2a (1.12)
I /r
t 2 dt = 3 (t2 + t2t1 + ti) (1.15)
t2 - tl Jtl
which gives
at = (t2 - t1) (1.16)
For this case the signal unambiguously lasts (t2 - tj ). However, 2ot is pretty close to
the true duration and has the advantage that it applies to any signal.
6 Chap. 1 The Time and Frequency Description of Signals
related we may think of the spectrum as the signal in the frequency domain or fre-
quency space or frequency representation.
Spectral Amplitude and Phase. As with the signal, it is often advantageous to write
the spectrum in terms of its amplitude and phase,
We call B(w) the spectral amplitude and 7&(w) the spectral phase to differentiate them
from the phase and amplitude of the signal.
Energy Density Spectrum. In analogy with the time waveform we can take I S(w) 12
to be the energy density per unit frequency:
That I S(w) 12 is the energy density can be seen by considering the simple case of
one component, s(t) = S(wo) e310t, characterized by the frequency, wo. Since the
signal energy is I s(t)12, then for this case the energy density is I S(wo)12. Since all
the energy is in one frequency, I S(wo) 12 must then be the energy for that frequency.
In Chapter 15 we consider arbitrary representations and discuss this issue in greater
detail. Also, the fact that the total energy of the signal is given by the integration of
I S(w) 12
over all frequencies, as discussed below, is another indication that it is the
density in frequency.
The total energy of the signal should be independent of the method used to
calculate it. Hence, if the energy density per unit frequency is I S(w) 12, the total
energy should be the integral of I S(w) 12 over all frequencies and should equal the
total energy of the signal calculated directly from the time waveform
(1.23)
J
where in going from Eq. (1.21) to (1.22) we have used
1 ej("-"')t
dt = 6(w - w') (1.24)
27r
(w) = I (1.25)
J
B2=c,22 = J(w_(w))2 IS(w)I2dw (1.26)
What Does the Energy Density Spectrum Tell Us? The energy density spectrum
tells us which frequencies existed during the total duration of the signal. It gives us
no indication as to when these frequencies existed. The mathematical and physical
ideas needed to understand and describe how the frequencies are changing in time
is the subject of this book.
discuss the method in the context of its significant calculational merit: To calculate
averages of frequency functions we do not have to calculate the Fourier transform,
S(w). It can be done directly from the signal and done simply. We first state the
result for average frequency and average square frequency, give a few examples to
display the power of the method, and then discuss the general case.
The average frequency and average square frequency are given by
z
(w2) = fw2 I S(w) Iz dw = Js*(t)
(j dt) s(t) dt (1.30)
J
s* (t) d s(t) dt
Zt (1.31)
(1.32)
dt s(t) Idt
That is, to calculate the average frequency, we differentiate the waveform and carry
out the integration as per Eq. (1.29). For the average square frequency we have a
choice. We can calculate the second derivative and use Eq. (1.31) or calculate just
the first derivative and use Eq. (1.32). Either way we get the same answer. The
bandwidth is given by
Qz
w f (w - (w))2 I S(w) 12dw (1.33)
r lz
- (w )J s(t) dt
1d2
s*(t) (1.34)
J (-_
j dt - (w )) s(t) dt (1.35)
These results are far from obvious, although they are easy to prove and we will do
so shortly. First, a few examples.
Take
3(t) = (a/,n.)1/4 e-ct2/2+jOt2/2+3 0t
Now
s(t) _ { jat +,at + wo J '9(t) (1.37)
3 dt
and therefore
10 Chap. 1 The Time and Frequency Description of Signals
2
s` (t) { jut +,3t + wo } s(t) dt = { jat + fit + wo } e-t dt = wo
J
(1.38)
Also,
which gives
2 a2 q2
+
a. = (1.40)
2a
For this case the spectrum and energy density spectrum may be obtained without
difficulty,
S(w) _ a
7( - jQ)
e-
(_-_0)2/2(&-j#)
(1.41)
/(Q2+9')
S(w) I2 = 7r(a2 + Q2))
(1.42)
This example is chosen to impress the reader with the simplicity and power of the
method. Suppose we have the signal
s(t) = (1.43)
The hard way: The reader is welcome to try to calculate the Fourier transform of s(t)
and then use it to calculate the average frequency and bandwidth.
For the average square frequency we immediately get, using Eq. (1.32),
z
(w2) = at + j,3t + jMW",, COsW..t + jwo 12 a-, dt (1.48)
7
This integral is easy to evaluate since all the terms in the integrand that are linear in
time drop out due to symmetry. All complex terms drop out also since we know that
the answer has to be real; therefore the complex terms must add up to zero and do not
have to be calculated. The remaining terms are simple integrals. Evaluation leads to
a 2a3 2 2
(1+
(w2) +wp
m e-W'J (1.49)
2
which gives
012 = a
2 2
+ M2 2' ` (1
-
2
Q e--'/2)
(1.50)
2 2
The Frequency Operator and the General Case. For convenience one defines the
frequency operator by
1d
(1.51)
3 dt
and it is understood that repeated use, denoted by Wn, is to mean repeated differ-
entiation,
(1)7' n
Wns(t) _ 3(t) (1.52)
dtn
We are now in a position to state and prove the general result that the average of a
frequency function can be calculated directly from the signal by way of
In words: Take the function g(w) and replace the ordinary variable w by the operator
dt; operate on the signal, multiply by the complex conjugate signal, and integrate.
Before we prove this we must face a side issue and discuss the meaning of g(W) for
an arbitrary function. If g is wn, then the procedure is dear, as indicated by Eq. (1.52).
If g is the sum of powers, then it is also dear. For a general function we first expand
the function in a Taylor series and then substitute the operator W for w. That is,
To prove the general result, Eq. (1.54), we first prove it for the case of the average
frequency:
Jsl(t).s(t)dt (1.60)
These steps can be repeated as often as necessary to prove it for g = w'. Hence
w s t
()fl s(t dt= f ?tW st dt (1.61)
Having proved the general result for functions of the form g = w', we now
prove it for an arbitrary function, g(w), by expanding the function in a Taylor series
fs(t)(W')s(t)dt (1.64)
This is readily proved by integrating by parts. Also, a real function, g(w), of a Her-
mitian operator, g(W), is also Hermitian. That is,
The Time Operator. In the above discussion we emphasized that we can avoid the
necessity of calculating the spectrum for the calculation of averages of frequency
functions. Similarly, if we have a spectrum and want to calculate time averages, we
can avoid the calculation of the signal. The time operator is defined by
T = -i du) (1.70)
In particular,
&A)2
S(w) dw (1.73)
S* S dw (1.74)
I d
dw
S(w)
2
dw (1.75)
14 Chap. 1 The Time and Frequency Description of Signals
We want to calculate the mean time and duration directly from the spectrum. We have
and therefore
12
[ to - ja/2 + jn/w ] I S(w) dw = to (1.78)
1
In evaluating this we don't have to do any algebra. The complex terms must be zero
and hence need no evaluation. Also, since the spectrum is normalized, the real part
integrates to to. Similarly
which gives
T2 = (t2) - (t)2 = 1 a2
(1.80)
4 2n - 1
The signal corresponding to this spectrum is
a2n+1 1
s(t) = n! (1.81)
27r(2n)! [ a/2 - j(t - to) ]n+1
and the above average may be checked directly. Also, for this signal
(w) - 2n + 1 2 (2n+2)(2n+1) 2 2n+1
(w ) = a2 B - a2 (1.82)
a +
The Translation Operator. Many results in signal analysis are easily derived by the
use of the translation operator, e?TN', where -r is a constant. Its effect on a function
of time is
e'Twf(t) (1.83)
= f(t+T)
That is, the translation operator translates functions by z. Note that it is not Hermi-
tian. To prove Eq. (1.83) consider
n=0
f (t) = E n! d f (t)
n=0
(1.84)
But this is precisely the Taylor expansion of f (t + T) and hence Eq. (1.83) follows.
Similarly, the operator a-,j9T translates frequency functions,
e-i9T S(w) = S(w + 0) (1.85)
15
Sec. 5 Bandwidth Equation
The Operator Method. We have shown how the properties of the time and fre-
quency operators can be used to simplify calculations. Indeed, it is a powerful
method for that purpose of which, hopefully by now, the reader is convinced. How-
ever, the operator method is not only a calculational tool but one of fundamen-
tal significance which will be developed and discussed in the later chapters of this
book
The second term is zero. This can be seen in two ways. First, since that term is
purely imaginary it must be zero for (w) to be real. Alternatively, we note that the
integrand of the second term is a perfect differential that integrates to zero. Hence
This is an interesting and important result because it says that the average fre-
quency may be obtained by integrating "something" with the density over all time.
This something must be the instantaneous value of the quantity for which we are
calculating the average. In this case the something is the derivative of the phase,
which may be appropriately called the frequency at each time or the instantaneous
frequency, wz(t),
wt(t) _ P (t) (1.91)
Instantaneous frequency, as an empirical phenomenon, is experienced daily as
changing colors, changing pitch, etc. Whether or not the derivative of the phase
meets our intuitive concept of instantaneous frequency is a central issue and is ad-
dressed in subsequent chapters. In addition, this brings up the question that if
16 Chap. 1 The Time and Frequency Description of Signals
instantaneous frequency is the derivative of the phase, what phase are we to use?
According to this definition the instantaneous frequency of a real signal is zero,
which is clearly an absurd result. The means to get around these difficulties are
developed in the next chapter with the introduction of the complex signal which
corresponds to the real signal.
B2 0-z
U) f (w-(w))z IS(w)I2dw (1.92)
z
fs*(t) (- dt - (w) s(t) dt (1.93)
r 2
ld
J \ j dt - (w 8(t) dt (1.94)
I 1
A(t) + 0'(t) - (w ) A2(t) dt (1.95)
or
B2
r(:? z
A2(t) dt + (cp'(t) -(w})z A2(t)dt (1.%)
=J )
J
While A2(t) can be canceled in the first term, the expression as written is prefer-
able because it explicitly shows that the bandwidth is the average of two terms, one
depending on the amplitude and the other depending only on the phase. What is
the meaning of this formula? What is the significance of the two terms? The expla-
nation will come with the ideas of describing a signal jointly in time and frequency,
which is the topic of this book The same steps lead to
Consider
e-,,t2/2+.7t3/3+j8t'/2+j.ot
s(t) = ((Y/7r)1/4 (1.99)
Sec. 6 AM and FM Contributions to the Bandwidth 17
The denvative of the phase is given by gyp' = wo + (3t + -yt2 and hence
For the average square frequency we note that A'/A = -at and therefore
3,Y2
23 2
+ w + 4a2 + yao (1.102)
2a z
from which we obtain
a2+o2
2
U2 = 72 (1.103)
2a + 2a2
with
2
B2 =
M + BFM
BA2
(1.105)
We also define the fractional contributions and the ratio of the two contributions by
(a)
Or \- 0-
-,. .,.
0 5 0 5
(c`
D
1-
0 t -a 50 0 50
Fig. 1.1 The real part of the signal s(t) = (a/7r)1/a e-ate/2+jat2/2 for various val-
ues of a and 0. All signals have the same bandwidth. There is a large amplitude
modulation in (a) and a small amplitude modulation in (d) but large frequency de-
viations from the mean frequency. The AM and FM contributions as defined by Eq.
(1.104) are a measure of the two qualitatively different effects that are contributing
to the bandwidth. The AM contributions are 71, 18, 1.3 and 0.4 % respectively, as
we progress from (a) to (d). The values for a,,3 are (16,16 ), (1, 5.56), (0.005, 0.4), and
(0.0005, 0.127) and each pair gives a bandwidth of 4.
This is reasonable because a constant frequency does not contribute to the bandwidth,
since the bandwidth indicates the spread in frequency.
TAM =
a
a2 + rFM =
_ p _ (1 . 111)
N2 a2 + F' 2
Sec. 7 Duration and Mean Time in Terms of the Spectrum 19
A "large" bandwidth can be achieved in two qualitatively different ways. Since the
instantaneous frequency is w: = wo + fit, we can get a large spread by waiting long
enough. We achieve this by making the duration long, which is accomplished by tak-
ing a small a. However, we can also get a large bandwidth by making the duration
very small, that is by taking a large a.
BAM
2
R.., = '- m (1 - C-w,12n (1.113)
The instantaneous frequency is mw,,, cosw,,,t + wo and we see that we can indeed
achieve a large bandwidth if we have a large modulation index, m, or large w,,,, since
by making either one large we get a large instantaneous frequency change if we wait
long enough. As with the previous example we can achieve the same bandwidth by
making the duration small. Qualitatively we have two very different effects and the
AM and FM expressions measure these two effects.
and
T2
(Bp))2
= Ut = B2(w) Caw + f (,O,(w) + (t))2 B2(w) dw (1.115)
J B(w)
Examine Eq. (1.114). It says that if we average -0'(w) over all frequencies we
will get the average time. Therefore we may consider -i'(w) to be the average
time for a particular frequency. This is called the group delay and we shall use the
following notation for it
t9(w) = - "(w) (1.116)
In Section 1.6 we showed how amplitude and frequency modulation contribute
to the bandwidth. Similarly, the amplitude and phase variations of the spectrum
20 Chap. 1 The Time and Frequency Description of Signals
with
T 2 = TSAM + TSPM (1.118)
which we have given in Eq. (1.41) and which we rewrite here in terms of the spectral
phase and amplitude,
S(w)
= (a/7r)1/4 expp - Wo)2 i3(- )2 1
( 1 . 120 )
va - 2(a2 +,32) 2(a2 + 02) j
Therefore,
tg = a2
a p2 (w - WO) ( 1 . 121 )
Also,
2 a 2 $2
TSAM = 2(a2 + 02) TSPM = 2a(a2 + $2) (1 . 122)
The reason for defining the correlation coefficient in the standard considerations
such as for height and weight is that it ranges from minus one to one and hence
gives an absolute measure. That is not the case here, but nonetheless it does give a
good indication of the relationship between time and frequency.
Are these two definitions, Eq. (1.124) and Eq. (1.127), identical? For them to be
identical we must have
f t cp'(t) I s(t) l2
dt = - Jw '(w) I S(w) 12 dW (1.128)
In fact, this equation is true but not obvious. It can be proven by brute force, but
a very simple proof is given in Section 15.4. It is an interesting identity because it
connects the phases and amplitudes of the signal and spectrum.
and therefore the covariance and correlation coefficient are equal to zero. Similarly,
if we have a spectrum of the form S(w) = B(w)eiwto, then there is no correlation
22 Chap. 1 The Time and Frequency Description of Signals
or
S(w) = B(w) e3`''to (1.133)
Covariance of a Real Signal or Spectrum. Since the phase of a real signal is zero, the
derivative is zero and hence the covariance of a real signal is always zero. This result
misrepresents the physical situation and is another reason for defining a complex
signal as we discuss in the next chapter. Similarly, signals that are symmetric in time
have real spectra and their covariance is zero.
A
Cove,,, = 2 ; r= (1.136)
a _+022
When p --' 0 the correlation coefficient goes to zero, in conformity with our discussion
that for a constant frequency signal, time and frequency are not correlated. As a -+ 0
the correlation coefficient goes to 1, depending on the sign of Q. That is, for constant
amplitude we have total correlation. If f3 is positive, the instantaneous frequency in-
creases with increasing time and that is what the value of +1 for r tells us. There is
high positive correlation in the sense that when time is large, the value of the instan-
taneous frequency is also large. For Q negative we have negative correlation, which is
also reasonable, since as time is increasing the instantaneous frequency is decreasing.
For a - oo we have the correlation coefficient going to zero, which is also reason-
able because for that case we have a short duration signal and the frequencies bear no
relation to the chirping but are simply due to the short duration.
But we know from Section 1.4 that ej'w is the translation operator and therefore
Since this function compares or correlates the signal at two different times it is com-
monly called the (deterministic) autocorrelation function. Inversely, we have
ejet e,oTS(w)
dc, = f S"(w) S(w-e) dw (1.140)
R(9) = f I s(t)12 dt = f S*(w)
and hence
I s(t) 12 =
27r
f R(O) Cite d9 (1.141)
However, the energy density is not the sum of the energy densities of each part
1 S 12
= I S1 + 5212 = 15112 + 15212 + 2 Re { Si S2 }
0 IS112+IS212
Thus the frequency content is not the sum of the frequency content of each signal.
One cannot think of frequency content of signals as something we have in a bucket
that we add to the frequency content or bucket of another signal. The physical
reason is that when we add two signals, the waveforms may add and interfere in
all sort of ways to give different weights to the original frequencies. Mathematically
24 Chap. 1 The Time and Frequency Description of Signals
s(t)
Ft -
0
A
13 0 13 0 13 0 13
IS(U,)I'
J
0
W --r
1s 0 15 0 15 0
AN 15
Fig. 1.2 The energy density spectrum for the sum of two signals, each localized in
time. The signal is given by Eq. (1.145), where a, = ctiz, W1,2 = 5, 10, ar 1 t1 = 5.
The values for t2 are 5, 7, 9, and 11 in (a)-(d). Even for large time separations as in
(c) and (d) the energy density spectrum is affected. The density of frequency is not
just the sum of the frequencies of each part. The implications for time-frequency
analysis are explained in the next figure.
this is reflected by the fact that the energy density spectrum is the absolute square
of the sum of the spectra, which results in nonlinear effects. How the intensities
change is taken into account by Eq. (1.143).
Even if the signals are well localized and greatly separated in time we still cannot
add frequency content. For example, consider two signals each lasting a second but
one happening today and the other a million years from now. Suppose we add the
two signals. Because of the large separation in time we have the intuitive sense that
physical properties should somehow also be separable and simply add, but that is
not the case. Even though the two signals have in some sense nothing to do with
each other since they are greatly separated in time, nonetheless the sum of the two
signals produces an energy density spectrum that is not simply the sum of the two
energy density spectra. Let us take a particular example:
If we take eel and a2 to be large in relation to It2 - tt I the two parts will be well
separated in time. A few typical situations are illustrated in Fig. 1.2, where the en-
ergy density spectrum is also shown. The energy density spectrum changes with
signal separation even when the separation is large. This is an important consider-
ation since it leads to seemingly paradoxical results in the distribution of intensity
in the time-frequency plane. Even though we have not begun our discussion of
time-frequency distributions it is important to understand this nonadditivity prop-
erty. Suppose we have two signals, each concentrated around the time-frequency
points w1, tl and w2i t2 and where these two points are well separated in both time
and frequency. If we add these signals we may think that since they are well sepa-
rated in time and frequency the resulting time-frequency density would be the sum
of the same two dumps. However, that cannot be the case since it does not take into
account the term Si S2 + S1 S2 in the energy density spectrum. This is illustrated in
Fig. 1.3.
Classification of Signals 25
Sec. 11
(t2,w2)
(tz, w2)
(ti, wi)
0 (c)
Fig. 1.3 In (a) we have a signal that is localized at the time-frequency point (t2, w2)
and in (b) we have another signal localized at (tl,wl). Since (ti, wi) and (t2, W2)
are separated and the densities localized, a seemingly plausible representation of
the energy density in the time-frequency plane for the sum of the two signals of (a)
and (b) is shown in (c). However, this cannot be correct because that would imply
that the energy density spectrum is the sum of the density spectra of each signal,
which is not the case, as discussed in the previous figure.
ple, if a signal varies from 100 to 5000 Hz in 10 seconds in a pure and steady way,
then to classify it as broadband does not present a complete picture, since at each
time there may have been only one frequency. On the other hand we can produce
a signal that has the same bandwidth where indeed at each time there is a broad
range of frequencies. Time-frequency analysis enables us to classify signals with a
considerably greater reflection of the physical situation than can be achieved by the
spectrum alone.
Chapter 2
Instantaneous Frequency
and the Complex Signal
2.1 INTRODUCTION
Signals in nature are real. Nevertheless, it is often advantageous to define a complex
signal that in some sense or other corresponds to the real signal. In this chapter we
describe the motivations for seeking a complex signal representation and its relation
to the concept of instantaneous frequency. We have seen in Chapter 1 that it is
natural to define instantaneous frequency as the derivative of the phase because its
average over time is the average frequency. Thus far we have left open the question
of how to get the phase. One of the motives for defining the complex signal is that
it will allow us to define the phase, from which we can obtain the instantaneous
frequency.
We seek a complex signal, z(t), whose real part is the "real signal", sr(t), and
whose imaginary part, sli(t), is our choice, chosen to achieve a sensible physical
and mathematical description,
If we can fix the imaginary part we can then unambiguously define the amplitude
and phase by
which gives
(sisr - srsi) /A2 (2.3)
27
28 Chap. 2 Instantaneous Frequency and the Complex Signal
2 vw 2 or,,
(b)
IS(w)I2
0
w--4 (w)l
Fig. 2.1 A real signal has a symmetrical density in frequency, as illustrated in (a).
Therefore, the average frequency is zero and the spread in frequency is roughly the
distance between the two bumps, as illustrated in (a). Neither of these results is in-
dicative of the physical situation. The analytic signal is defined to give the identical
spectrum for the positive frequencies and zero for the negative frequencies, as in
(b). This results in an average frequency and bandwidth, which better reflect the
physical situation in that the average frequency falls somewhere in the middle of
the bump and the bandwidth is the spread of the bump.
on. We want the answer to come out to be somewhere in the middle of the right-
hand bump. Also, the spread in frequencies will be roughly the distance between
the two bumps, while what we want is the spread of one bump. What can we do to
obtain a value for average frequency that is roughly centered in the middle of the
right hand bump? We can achieve this by simply neglecting the left bump in the
averaging,
There are now two approaches we can take. First, we can continue to consider real
signals and when taking spectral averages integrate from zero to infinity rather than
-oo to oo. Or, we can define a new signal that has the same spectrum for the pos-
itive frequencies and a zero spectrum for the negative frequencies. The advantage
of the second approach is that we can calculate frequency averages directly from
the signal and therefore it is advantageous to have the signal once and for all. In
particular, the new signal, z(t), as yet unknown, will assure that
The second reason for wanting to form a complex signal is that it will allow us
to obtain the phase and amplitude of a signal unambiguously, and that allows us to
obtain an expression for instantaneous frequency.
30 Chap. 2 Instantaneous Frequency and the Complex Signal
The factor of 2 is inserted so that the real part of the analytic signal will be s(t);
otherwise it would be one half of that. We now obtain the explicit form for z(t) in
terms of the real signal s(t). Since
r 00
e_j"t
S(w) = 2 s(t) dt
J 0
we have
z(t) =2
j ,I s(t') a"' e'"t d' dw
= 1r r s(t') dt' dw
o f
and using
(2.10)
we obtain
z(t) _ 1. 18(t1) 17r 6(t - t') + t, ] dt' (2.11)
t
yielding
We use the notation A[s] to denote the analytic signal corresponding to the signal s.
The reason for the name analytic is that these types of complex functions satisfy the
Cauchy-Riemann conditions for differentiability and have been traditionally called
analytic functions. The second part of Eq. (2.12) is the Hilbert transform of the signal
and there are two conventions to denote the Hilbert transform of a function, s(t)
and H[s(t) ]
IIn a series of seminal papers Vakman 1W, 549, 5511 has addressed the concepts of instantaneous fre-
quency and the analytic signal and has brought forth the fundamental issues regarding these subjects.
The classical review article on the subject is by Vakman and Vainshtein.15501
Sec. 4 Calculating the Analytic Signal 31
The Analytic Signal of a Complex Signal. We have emphasized the physical mo-
tivation for defining a complex signal that corresponds to a real signal. However,
there is no reason why s(t) in Eq. (2.12) cannot be any function, real or complex.
For manipulative purposes it is important to allow for that possibility.
Energy of the Analytic Signal. Because we have insisted that the real part of the
complex signal be the original signal, normalization is not preserved. Recall that
the spectrum of the original real signal satisfies I S(w) I _ I S(-w) I and therefore
the energy of the original signal is
E. = f I S(w) I2 dw = 2
f 0
I S(w) I2 dw = 1
2 J0
12S(w) I2 dw = 2Ez (2.14)
That is, the energy of the analytic signal is twice the energy of the original signal.
In addition, the energy of the real part is equal to the energy of the imaginary part
Ea = EHfa] (2.15)
which can be seen by considering I z(t) 12 = I s(t) +jH[s] 12. When this is expanded
the middle term is
s` (t) s(t') + s(t)
ff t - t' 87 (t') dt' dt =0 (2.16)
A[ e3"t] 0 if w < 0
2 ei"t if > 0 (2.17)
32 Chap. 2 Instantaneous Frequency and the Complex Signal
This simple result is very important because if we can express the signal in terms of
exponentials then all we have to do to form the analytic signal is drop the terms with
negative frequency and multiply the positive frequency terms in the expansion by
two. A few examples will make this clear.
e;1-1t (2.21)
Similarly,
A[sin IwItj = 2,j A[e'IWlt -e-.iHWltj = -je;IWlt (2.22)
The last two terms have negative frequencies and hence only the first two terms re-
main, giving
z(t) = 24 (ei(W2+W1)t + e3(W2-W1)t) \\11
= 2
(eyWlt
+ e-7"lt) e7"2t (2.26)
= coswlte) 2t (2.27)
Notice that the analytic signal procedure chooses the higher frequency for the in-
stantaneous frequency.
The Analytic Signal of an Analytic Signal. If we start with an analytic signal then
its spectrum is nonzero only on the positive frequency axis and hence there is noth-
ing to drop. Therefore we get back the same signal except for the factor of two,
.Al z(t) i = 2z(t) [ if z(t) is analytic] (2.28)
Analytic Signal of the Derivative. Suppose we want the analytic signal of the
derivative of a function. If we have the analytic signal of the function then all we
Sec. 4 Calculating the Analytic Signal 33
have to do is differentiate it, because the analytic signal of the derivative of a func-
tion is the derivative of the analytic signal. More generally, the analytic signal of
the nth derivative of a function is the nth derivative of its analytic signal. To see this
consider
d'tis A do 1 S(w) e'"t
A (2.29)
dtn dtn 271 TO,
0
But this is just the nth derivative of the analytic signal of s(t). Therefore
1dns
A dtn ]
-A[s]
do
dtn
(2.33)
We note that in going from Eq. (2.30) to Eq. (2.31) we have used the fact the ana-
lytic signal operator is linear and hence can be taken inside the integration because
integration is a sum.
Convolution. Suppose we have a signal, s(t), whose spectrum is S(w) and we form
a new signal by way of
where F(w) is any function. It is dear that z(t) is analytic since we have made it so
by construction. By specializing the function F(w) many important results can be
obtained. In particular, if we consider F(w) to be the spectrum of a function, f (t),
and recall that the product of two spectra is the convolution of the two signals, then
we have in general that
That is, the convolution of an analytic signal with an arbitrary function results in
an analytic signal.
hertz, is raised to the megahertz range because we can build transmitters and re-
ceivers at those frequencies and because electromagnetic waves propagate well in
"0' at those frequencies. What is done is to multiply the original wave-
our atmosphere
form by e) with wo positive to form a new signal,
snew(t) = 3(t) e'--t [ wo > 0 ] (2.36)
In such a case one says that e3"t is the imposed modulation and wo is the carrier
frequency. The spectrum of the new signal is the spectrum of the old signal shifted
by an amount wo
Snew(w) = S(w - wo) (2.37)
When will the resulting signal be analytic? If the spectrum of the original signal
extends from minus infinity to infinity then the spectrum Sne,,, will likewise extend
from minus infinity to infinity and not be analytic. But suppose the spectrum S(w)
is single sided, which means that it is zero below a certain value. Then if wo is greater
than that value, the spectrum of the new signal will be shifted to the positive part
of the frequency axis and the signal will be analytic. Therefore
s(t) e3"0t is analytic if the spectrum of s(t) is zero for w < -WO (2.38)
Note that we have not specified whether s(t), the original signal, is real or complex.
It can be complex. If it is real then this result implies that it must be bandlimited to
the interval (-wo,wo).
Analytic Signal of the Sum of Two Signals. Suppose we have a signal and want to
add to it another signal so that the resulting sum will be analytic. All we have to do
is choose a signal whose spectrum for negative frequencies is identical to the first
signal but differs in sign. That way, the spectrum of the resulting signal will be zero
for negative frequencies. Therefore
exponential with the worst case of si will result in a negative frequency. Therefore
s2 cannot have any frequencies below wl and hence
A[ 81821 = s1A[ 821 if the spectrum of s1 is zero below the value -w1 and s2 is
any signal whose analytic function has a spectrum that
is zero below w1. (2.40)
Notice that the spectrum of s2 does not have to be zero below w1; it is the spectrum
of A[ 521 that has to be zero below w1. If 52 is real, then the condition is that the
spectrum of s2 must be zero outside the range (-wi, wi ). These results are due to
Bedrosianl641 and Nuttall.1402]
Product of Two Analytic Signals. If si and 82 are both analytic, then they are both
single sided and we have
A[81521 = s1A[ s2 ] = 2s132 sl and 82 analytic (2.41)
Real Signals. Suppose si and 82 are real. Their spectra are then symmetric. Since
the spectrum of Si vanishes below -wl it must vanish above wi. Therefore the
spectrum of s1 must be zero for Iw > wl. Similarly, the spectrum of s2 must vanish
for [ w i < wi. Hence
A[ sis2 ] = s1A[ s2 ] for real signal if spectrum of si is zero for jwj > wi (2.42)
We now ask what the analytic signal procedure has done in terms of choosing the
particular amplitude and phase, that is, what is special about the amplitude and
phase to make it an analytic signal? Generally speaking, the answer is that the
spectral content of the amplitude is lower than the spectral content of e.iw(t). We
illustrate this first with a simple case where
SA(W) =
27r
r A(t) e-i"'t dt (2.45)
36 Chap. 2 Instantaneous Frequency and the Complex Signal
The spectrum of s(t) is then SA(W - wo). For SA(W - wo) to be analytic, that is, zero
for negative frequencies, SA (w), it must be zero w < wo. Therefore
A(t) e3 0t is analytic
if the spectrum of A(t) is contained within (-wo, w0) (2.46)
Hence all the low frequencies are in the amplitude, and the high frequency is in the
cosine. One can generalize this result in the following way. Call S,e (w) the spectrum
of ej`e(t),
S,(W) = 1 eaw(t)
(2.47)
27r
S(W)= 1
vlr-
2=
f SA(w-w')Sw(w')dw' (2.48)
We can consider this the sum of shifted spectra of A(t) with coefficients S,(w').
Now suppose SA(w) is bandlimited in the interval (-w1, wi). A sufficient condition
to shift S(w) to the positive axis is if the lowest value of the range of w' is greater
than wl. That is, S, (w') is zero for values less than wl,
Therefore what the analytic procedure does, at least for signals that result in the
above forms, is to put the low frequency content in the amplitude and the high
frequency content in the term ej`e(t).
at least for some situations. The sq(t) thus formed is called the quadrature model
signal. This idea was used before the introduction of the analytic signal. As we
mentioned in the introduction, it begs the question since the procedure does not
tell us how to write the signal in the form A(t) cos W (t) to start with. Nonetheless,
in many situations we may think we know the phase and amplitude. Or we may
have an A(t) and ap(t) and want to construct a signal and ask whether or not it is
analytic. More importantly, since calculating the analytic signal is difficult, if we
Sec. 6 The Quadrature Approximation 37
Energy Criteria. The less of the spectrum of the quadrature signal there is on the
negative frequency axis, the better is the agreement. To examine this question more
precisely we define the following three signals:
sa(t) =
2
27r
f
o
S(w) e'"'t dw analytic signal (2.53)
where we have used the fact that S(w) = S*(-w), since s(t) is real. Also, the spec-
trum of the analytic signal is
Sa(w) 0 ifw<0
(2.57)
2 S(w) = Sq (w) + Sq (-w) if w > 0
As a criterion on closeness between the quadrature signal and analytic signal,
Nuttall(''02) considered the energy of the difference between the two signals,
of = f I
(2.58)
fpsq(w)2th+j!S;(_w)i2th
o
100
(2.60)
38 Chap. 2 Instantaneous Frequency and the Complex Signal
which is twice the energy of the quadrature model in the negative part of the spec-
trum.
00
(2.62)
- 1
27r J
0
00
Sq(w) e& "t dw + 1
27r
1- S(-) eJ"t dw (2.63)
=
1
r a-j"t - Sq (w) "t I dw (2.64)
2qr J-cc [ Sq (w)
This compares the signal at every time. If we take absolute values, then
I sa (t) - sq (t) I = I
1
2qr f
00 [ Sq (w)
a-j"t - Sq (w) &"t ] dw (2.65)
Imposed Amplitude and Phase Modulation. Often we have a carrier and impose
an amplitude and phase, A(t) and cp(t), to produce the complex signal
s(t) = A(t) ej"t+.iw(t)
(2.68)
The spectrum of A(t) ejw(t) is shifted upward by an amount w0 and therefore, using
Eq. (2.61),
r0
DE = 2 1 Sq (w - w0) 1 2 dw (2.69)
J 00
2-ISq(w)12dw (2.70)
f
Note that Sq(w) is still the spectrum of A(t) ej`e(t).
Sec. 7 Instantaneous Frequency 39
How does this compare with the analytic signal of s(t)? Using Eq. (1.42) for the spec-
trum of sq we have
a]2
/ o
a+A
AE=2J IS(w)12dw=1-. a-12dx (2.73)
00 fo
This will be close to zero when the upper limit is close to infinity and therefore
where QW is the bandwidth of sq(t) as per Eq. (1.40). This is reasonable because for
a given wo, the energy density spectrum will not spill over much into the negative
frequency region if it is relatively narrow.
Paradoxes Regarding the Analytic Signal. There are five paradoxes or difficulties
regarding the notion of instantaneous frequency if it is defined as the derivative
of the phase of the analytic signal. To some extent we will be able to understand
and resolve some of these paradoxes when we study the concept of instantaneous
bandwidth in Section 13.2 and with the introduction of the idea that instantaneous
frequency is a conditional average frequency. It is important to understand these
difficulties in the most basic terms because in more complicated situations their re-
flection may cause difficulties in interpretation.
First, instantaneous frequency may not be one of the frequencies in the spec-
trum. That is strange because if instantaneous frequency is an indication of the
frequencies that exist at each time, how can it not exist when we do the final book-
keeping by way of the spectrum? Second, if we have a line spectrum consisting of
only a few sharp frequencies, then the instantaneous frequency may be continuous
and range over an infinite number of values. Third, although the spectrum of the
analytic signal is zero for negative frequencies, the instantaneous frequency may
be negative. Fourth, for a bandlimited signal the instantaneous frequency may go
outside the band. All these points are illustrated by the following simple example.
Consider
where the amplitudes Al and A2 are taken to be constants and wi and W2 are positive.
The spectrum of this signal consists of two delta functions at wi and w2,
Since we take wi and w2 to be positive, the signal is analytic. Solving for the phase and
amplitude,
Al sin wit + A2 sinw2t
ap(t) = arctan Ai cos wit + A2 cos wet (2.79)
A2 A21
wt = cps (t) = 2 (W2 + wl) + 2 (W2 - wi) A2 (t)
(2.81)
By taking different values of the amplitudes and frequency we can illustrate the points
above. This is done in Fig. 2.2.
Sec. 8 Density of Instantaneous Frequency 41
t --
0 t 2 3 4 0 1 2 3 4
Fig. 2.2 The instantaneous frequency for the signal s(t) = Ai e3lot + A2 &20t
The spectrum consists of two frequencies, at w = 10 and w = 20. In (a) Al = .2
and A2 = 1. The instantaneous frequency is continuous and ranges outside the
bandwidth. In (b) A, = -1.2 and A2 = 1. Although the signal is analytic the
instantaneous frequency may become negative.
This says that for a given w2, choose only the values for which wi = cp'(t). If there
is only one such value then this simplifies to
P(wi) S(t) )2
= (2.83)
I
4-t,
42 Chap. 2 Instantaneous Frequency and the Complex Signal
0'(t)
-(w))2Is(t)12dt (2.85)
This is precisely the first term appearing in the bandwidth equation, Eq. (1.%), and
therefore we can write that equation as
2
B2 = A'(t)
UIF +
f (A(t)) A2(t) dt (2.86)
But the second term is positive and hence we conclude that the spread in instanta-
neous frequency is always smaller than the bandwidth,["']
This may seem paradoxical since in the example above we showed that the spread
of instantaneous frequency can be outside the range of the frequencies that exist in
the spectrum. What this must mean is that while the instantaneous frequency can
range widely, the occurrences when it ranges outside the bandwidth of the signal
have small duration and hence do not contribute significantly to QIF. Note that the
spreads are equal when the amplitude of the signal is constant.
(2.88)
st (2.89)
P(wi) = 'W"(t) 2
t=(W.-WO)/#
_ a1 e-a(W;-W0)2/A2
(2.90)
V 7r Q
a e-a(W-W0)2/(a2+)92)
i S(w) 12 = (2.91)
C(a2 Q2)
Sec. 8 Density of Instantaneous Frequency 43
The spread of the spectrum, that is, the bandwidth, and the spread of the instanta-
neous frequency density are respectively
[a2 + Q2
Q- = arF = (2.92)
2c, 2
3.1 INTRODUCTION' 2
The time-bandwidth product theorem, or uncertainty principle, is a fundamental
statement regarding Fourier transform pairs. We are going to be particularly care-
ful in our discussion since the uncertainty principle has played a prominent role
in discussions, metaphysical and otherwise, of joint time-frequency analysis. The
discovery of the uncertainty principle in physics and chemistry is one of the great
achievements of the century. Unfortunately, it has generated many pseudo ideas,
which are rivaled only by the number of pseudo ideas generated by relativity, the
other great discovery of this century. The pseudo ideas in relativity are perhaps
characterized by "everything is relative ", and the uncertainty principle by "every-
thing is uncertain", neither view being remotely related to the physical or mathe-
matical truth.
'The uncertainty principle was first derived by W Heisenbergin the paper"Uber den anschhaufichen
Inhalt der quantentheoretichen Kinematic and Mechanik" ("On the Conceptual Content of Quantum
Theoretical Kinematics and Mechanics") in 1927. In that paper he derived the uncertainty relations on
the basis of a thought experiment involving resolution of an idealized microscope. In fact, Heisenberg
presented it as an equality rather than as an inequality. It was Weyl who subsequently saw that un-
certainty can be defined by the standard deviation and gave the proof commonly used today which
is based on the Schwarz inequality. The physicist C. G. Darwin (grandson of Charles Darwin) made
the connection between the uncertainty principle and Fourier transform pairs. E. U. Condon and H. P.
Robertson extended the uncertainty principle for arbitrary variables. In 1930 Schrbdinger grasped the
full generality of the uncertainty principle, derived it for arbitrary variables and saw the fundamental
connection with the commutator and anticommutator (see Eq. (15.87) ). While the mathematics of the
uncertainty principle was settled within five years of the discovery of quantum mechanics (1925-1926) by
Schrbdinger, Heisenberg, Born, and Dirac and within three years of Heisenberg's first paper on the sub-
ject, the interpretation and consequences of the uncertainty principle in quantum mechanics remains
a subject of great interest and activity and the number of papers written on it in the intervening sixty
years or so is truly phenomenal. After all, it is one of the most important discoveries of mankind.
2The uncertainty principle is discussed further in Sections 6.9 and 15.5.
44
Sec. 1 Introduction 45
For signal analysis, the meaning of the uncertainty principle and its importance
have been clearly stated often enough, although a mystery still persists for some.
Right on the mark is the statement by Skolnik[514): "The use of the word 'uncer-
tainty' is a misnomer, for there is nothing uncertain about the 'uncertainty rela-
tion'.... It states the well-known mathematical fact that a narrow waveform yields
a wide spectrum and a wide waveform yields a narrow spectrum and both the time
waveform and frequency spectrum cannot be made arbitrarily small simultane-
ously." Equally clear is Lerneri331:: The uncertainty principle " ... has tempted
some individuals to draw unwarranted parallels to the uncertainty principle in
quantum mechanics.... The analogy is formal only." AckroydJ41 has emphasized
that "There is a misconception that it is not possible to measure the t - f energy
density of a given waveform and that this is a consequence of Gabor's uncertainty
relation. However, the uncertainty principle of waveform analysis is not concerned
with the measurement of t - f energy density distributions: instead it states that if
the effective bandwidth of a signal is W then the effective duration cannot be less
than about 1/W (and conversely) ... "
Although we can hardly improve on the above we try. Here is the point: The
density in time is I s(t) 12 and the density in frequency is IS(w) I2; but s(t) and S(w)
are related and hence we should not be surprised to find that there is a relation
between the densities. The relation is such that if one density is narrow then the
other is broad. That's it and no more. It is not that both time and frequency cannot
arbitrarily be made narrow, but that the densities of time and frequency cannot both
be made narrow.
If the interpretation is so straightforward in signal analysis, why is it one of the
most profound discoveries in physics? Here is why: In classical physics and in every
day life it seemed clear that we can choose the position and velocity of objects at
will. No one imagined that we cannot place a ball at a given spot with a given
velocity. However, present day physics, quantum mechanics, says precisely that
and it is one of the great discoveries regarding the behavior of matter.
Quantum mechanics is inherently probabilistic. When we speak of densities in
quantum mechanics we mean probability densities and the word "uncertainty" is
appropriate because we are dealing with probability. In signal analysis the word
"uncertainty" is highly misleading.
Let us be very clear that in both physics and signal analysis the uncertainty prin-
ciple never applies to a single variable. It is always a statement about two variables.
Furthermore, it does not apply to any two variables, but only to variables whose
associated operators do not commute. In this chapter we deal with the uncertainty
principle for time and frequency and in Section 15.5 we consider its generalization
to other variables. There is a very elegant and simple way to derive the uncertainty
principle for arbitrary quantities and this is done in Section 15.5. Here we use a
simple brute force approach.
One of the reasons that there has been considerable confusion about the uncer-
tainty principle in signal analysis is that one very often modifies a signal, e.g. filters
it, windows it, etc. Once that is done we have two different signals, the original and
the modified one. Therefore we have two uncertainty principles, one relating to the
46 Chap. 3 The Uncertainty Principle
original signal and the other to the modified signal. Very often these are confused.
In Section 3.4 we will be particularly careful in making this distinction by deriving
the uncertainty principle for modified signals.
T2 = 0,i
= f(t - (t))2 I s(t) 2
dt (3.1)
B2 = a2 = f(w_(w))2IS()l2dw (3.2)
We emphasize that T and B are standard deviations defined in the usual manner
and no more. The uncertainty principle is
TB > z (3.3)
Therefore one cannot have or construct a signal for which both T and B are arbi-
trarily small.
What Does the Proof of the Uncertainty Principle Depend On? It is important to
have a clear picture of what the proof depends on so that no misleading interpre-
tations creep in. The proof depends on only four things: first, on I s(t) 12 being
the density in time; second, on taking I S(w) 12 as the density in frequency; third
that .9 (t) and S(w) are Fourier pairs; and fourth, on defining T and B as standard
deviations of time and frequency.
Notation. Very often the notation used to write the uncertainty principle is OtEw >
1. There is nothing wrong with this notation as long we understand that A means
standard deviation and nothing more. However, because A is typically used for the
differential element of the calculus or to signify "error" there is a tendency to think
of the uncertainty principle as having something to do with differential elements,
smallness, measurement, or resolution. The A of the uncertainty principle means
only one thing: the standard deviation. If this is kept in mind then no difficulties
arise in interpretation or philosophy.
Sec. 3 Proof of the Uncertainty Principle 47
has the same shape both in time and frequency as sold except that it has been trans-
lated in time and frequency so that the means are zero.3 Conversely, if we have
a signal s,, ,,,(t) that has zero mean time and zero mean frequency and we want a
signal of the same shape but with particular mean time and frequency, then
&('O)t
sold(t) = snew(t - (t))
The duration is
at!
= ft2s(t)I2dt (3.8)
and therefore
Qt QW = f t s(t) I2 dt x f I W(t) I2 dt (3.9)
Equation (3.9) is it; no other assumptions or ideas are used. The fact that s and S
are Fourier transform pairs is reflected in Eq. (3.7).
Now, for any two functions (not only Fourier transform pairs)
f I f(x) I2 dx f - ff*(x)g(x)dx
dx >
I z
(3.10)
which is commonly known as the Schwarz inequality.4 Taking f = is and g = a'
gives
at QU
z 2
f t s`(t)s'(t) dt
2
(3.11)
'The constant phase factor e< w) ( t) is irrelevantcan and be left out or in for the sake of symmetry.
4There are many proofs of the inequality. A simple one is to note that for any two functions
f f(.) 12 dx f ff*(z)9(x)dx z
fff(z)g(y)_f(y)g(x)I2dxdy
which is readily verified by direct expansion of the right hand side. Since the right hand side is manifestly
positive, we have Eq. (3.10).
48 Chap. 3 The Uncertainty Principle
= 2 d tA2 - 'A 2
+ j t 01 (t) (3.13)
The first term is a perfect differential and integrates to zero. The second term gives
one half since we assume the signal is normalized and the third term gives j times
the covariance of the signal. Hence
2
O_t2
Or2
f ts`(t)s'(t)dt =I-2+j Covt' 12 = 4 + Covt2,,, (3.14)
Therefore we have the uncertainty principle as given by Eq. (3.4). Since Cov W is
always positive, it can, if we so choose, be dropped to obtain the more usual form,
Eq. (3.3).
Minimum Uncertainty Product Signals. Since the minimum value for the uncer-
tainty product is one half, we can ask what signals have that minimum value. The
Schwarz inequality becomes an equality when the two functions are proportional
to each other. Hence, we take g = -cf, where c is a constant and the -1 has been
inserted for convenience. We therefore have
- c t s(t) = s' (t) (3.15)
This is a necessary condition for the uncertainty product to be the minimum. But
it is not sufficient since we must also have the covariance equal to zero, because by
Eq. (3.4) we see that is the only way we can actually get the value of 1/2. Since c is
arbitrary we can write it in terms of its real and complex parts, c = c,. + jcc. The
solution of Eq. (3.15) is hence
s(t) a-ct2/2 = e-(a,+ic:)t2/2 (3.16)
The covariance is the average value of t multiplied by the derivative of the phase.
The derivative of the phase is -c,t and, remembering that we are considering a
signal whose time and frequency means are zero, we have
r
ft cit1s(t)12dt-(w)(t)_-ctJ tee-'t2 (3.17)
The only way this can be zero is if c, is equal to zero and hence c must be a real
number. If we take c = a/2 we then have
(a/7x)1/4 e-at2/2
s(t) = (3.18)
This is the most general signal that has a time-bandwidth product equal to one half.
Sec. 3 Proof of the Uncertainty Principle 49
are given by Eq. (1.12) and (1.40). Using those values gives
at Qw
=2 a22p2
For this case the stronger version of the uncertainty principle yields an equality.
that
the spectrum is
a2ni 1 1
(3.26)
27r(2n)! [-a/2+j(w - wo) ]n+1
and
(t)- 2n+1 (t2) = (2n + 2)(2n + 1) 0'2 = 2n + 1 (3.27)
a a2 t a2
(w) = wo (wz) z a
= wo+2n-1
z
2 az
;
"' 2n-1 (3.28)
Therefore
1 2n + 1
at ow= (3.29)
2 2n-1
50 Chap. 3 The Uncertainty Principle
where h(t) is the window function, t is the fixed time for which we are interested,
and ,r is now the running time. This normalization ensures that
f I7)t(T) 12 dr = 1 (3.31)
for any t. Now 77i (T) as a function of the timer is of short duration since presumably
we have chosen a window function to make it so. The time, t, acts as a parameter.
The Fourier transform of the small piece of the signal, the modified signal, is5
Ft(w) = 1
e-j'T7)t(r)dT (3.32)
27r
51f the denominator in Eq. (3.30) is omitted then Ft(w) would become what is traditionally called
the short-time Fourier transform. In calculating quantities such as conditional averages, the normal-
ization must come in at some point and it is a matter of convenience as to when to take it into
account In Chapter 7 we use the more conventional definition and omit it from the definition.
There we use St (w) to denote the short-time Fourier transform and the relation between the two is
St (w) = Ft(w)I f 18(T) h(T - t) I2 dT}1/2.
Sec. 4 The Uncertainty Principle for the Short-Time Fourier Transform 51
Ft(w) gives us an indication of the spectral content at the time t. For the modified
signal we can define all the relevant quantities such as mean time, duration, and
bandwidth in the standard way, but they will be time dependent. The mean time
and duration for the modified signal are
frls(T)h(r-t)I2dr
(T )t =
f Tl77t(T)I 'd T = f Is(T)h(T-t)I2dT
(3 . 33)
T2 = (T - (T )) 2 77t(T)
2 dr = f (T f( I )2
h dT
(3 . 34)
J
I
3( h( 7T - t ) I2 dT)I2
Similarly, the mean frequency and bandwidth for the modified signal are
Bi = J(w_(w)t)2IFt(w)2dw (3.36)
This is the uncertainty principle for the short-time Fourier transform. It is a function
of time, the signal, and the window. It should not be confused with the uncertainty
principle applied to the signal. It is important to understand this uncertainty prin-
ciple because it places limits on the technique of the short-time Fourier transform
procedure. However, it places no constraints on the original signal.
We know that for infinitely short duration signals the bandwidth becomes in-
finite. Hence we expect that Bt oo as we narrow the window, which is indeed
the case. This is shown in Chapter 7 where we obtain explicit expressions for the
above quantities in terms of the amplitude and phases of the signal and window.
The point that we are making here is that we must be very clear with regard to the
uncertainty principle for the original signal and modified signal. It is true that if
we modify the signal by the technique of the short-time Fourier transform we limit
our abilities in terms of resolution and so forth. This is a limitation of the technique.
The uncertainty principle of the original signal does not change because we have
decided to modify it by windowing.
77, (T) _
/a+al1/4 exp _!(a+a)T2+aTt+jWor+jpT2/2- az t2
1\ a /1 a+a ,
(3.40)
from which we get
a 1
(T)` = a + a t T
T? =
- (3.41)
2(a + a)
z
at + WO Bt = (a + a) + 2 (3.42)
aa z a+ a
and therefore
z
Bt Tt = 2 1+ (a )2 (3.43)
For this case the local duration and bandwidth are time independent and so is the
uncertainty product. This is a peculiarity of the Gaussian window and will not be the
case with another window and/or signal. Now notice that as we narrow the window,
that is, as a -# oo, the bandwidth of the modified signal, Bt, goes to infinity. Of course,
this has nothing to do with the bandwidth of the original signal.
The mean time and durations are the same as in the preceding Example. The mean
frequency and bandwidth of the modified signal are
z
Mt 2 (a + a) + (a + a)2 t2
(3.45)
2 2
^f 2-y
BL = z (a + a) +
2(a + a)2 + (a + a)3
t2 (3.46)
Bt Tc
2
1 + (a +2
a)2 + (a + a)4
t2 (3.47)
For large times and short windows, the time dependent bandwidth goes as Bt
t2/a2; hence for large times or small windows we get a very large bandwidth. Again,
this has nothing to do with the original signal. The bandwidth of the original signal is
fixed.
Chapter 4
4.1 INTRODUCTION
Densities are the basic quantities we deal with, since we seek the energy density of
a signal at a particular time and frequency. We have already encountered densities
in the previous chapters, namely, the absolute square of the signal is the energy
density in time and the square magnitude of the Fourier transform is the energy
density per unit frequency. In this chapter we develop the basic methods and con-
cepts used to study and construct densities. We present the ideas in a manner that
is useful in studying energy densities in time, frequency, and joint time-frequency
densities.
By density we mean nothing more than the common usage, such as the density
of mass or the density of trees. Many of the ideas such as "standard deviation" will
have a probabilistic ring to them since they are usually associated with probability
densities. However, we emphasize that the methods presented in this chapter ap-
ply to any density. For the one dimensional case we should simply think of a wire
that may stretch from minus infinity to infinity and where the mass density varies
with location. Similarly, for the two dimensional case one should think of a sheet
where again the density varies with location on the sheet.
53
54 Chap. 4 Densities and Characteristic Functions
Since P(x)Ax is the amount in the interval Ax, the total amount is the sum,
Densities are often normalized so that the total amount is equal to one. The effect
of doing that, is, for example, if P(x) represents the actual number of people of a
certain height, then by normalizing it to one it represents the fraction of people at
that height.
Densities have to be single valued and positive. The positive requirement comes
from the fact that a density is the amount of something per unit something else
and both quantities are positive. It should be emphasized that the numerator is
the amount of something and while that something can be negative, the amount is
still positive. For example, while charge can be positive or negative the amount of
positive or negative charge is positive.
Averages. From the density simpler quantities are obtained which sometimes give
a gross indication of the basic characteristics of the density. The average, (x), is
The average of any function, f (x) can be obtained in two different ways: Either
from the density of x or from the density of u, P(u), where u = f (x),
How to obtain the density P(u) from P(x) is described in Section 4.7. Two general
properties of averages should be noted. First, the average of the sum of two func-
tions is the sum of the average of each function, (f (x) + g(x)) = (f (x) ) + (g(x) ).
Second, the average of a constant times a function is the constant times the average
of the function, ( cf (x) ) = c( f (x) ).
Standard Deviation. While the average gives the place where things are balanced,
it gives no indication as to whether the density is concentrated there or not. A
measure of that is the variance, which is the average of (x - (x ))2. Variance is
a good measure of concentration because if the density is concentrated near (x)
then (x - (x ) )2 will be relatively small. The weighting of that particular value is, of
course, taken with the density. The standard deviation, am, is defined as the square
root of the variance and given by
(x2)-(x)2 (4.7)
The significance of the moments are many fold. First, the first few moments give
an indication of the general properties of the distribution. Second, in some sense,
the more moments we know, the more we know about the distribution. Third,
for well behaved distributions the moments uniquely determine the distribution.
The procedure for constructing the distribution from the moments is presented in
the next section. In Chapter 10 we see how these methods can be used for the
construction of time-frequency distributions.
56 Chap. 4 Densities and Characteristic Functions
(xfz) (4.12)
which is a Taylor series in 9 with coefficients j" (x" ). Since the coefficients of a
Taylor series are given by the nth derivative of the function evaluated at zero, we
have
1 8"M(9)
(x") = in 59n (4.13)
B=o
The fact that moments can be calculated by differentiation rather than by integra-
tion is one of the advantages of the characteristic function, since differentiation is
always easier than integration. Of course, one has to first obtain the characteristic
function and that may be hard.
Fourier transform pairs are uniquely related and hence the characteristic func-
tion determines the distribution,
Taking the complex conjugate of Eq. (4.11) and using the fact that densities are real,
we have
M. (9) = f e-iGxP* (x) dx = M(-9) (4.16)
or
M-(-O) = M(O) (4.17)
The absolute value of the characteristic function is always less than or equal to one,
M(9) I < 1 (4.18)
Sec. 3 One Dimensional Characteristic Functions 57
I M(e) I =
f dx < f I ejex I I P(x) I dx = J P(x) dx = 1 (4.19)
We know that the characteristic function at the origin is equal to one and therefore
1 8M(9)
(x) _ 89
( -a20 + jxo) M(B) = x0, (4.23)
9=0 9=0
1 8M(8) 1 JA
= = 1/A
(x)
j 88 9=0 i (A-j6)z
9=0
(4.27)
z
(xz) = 1 82M(8) = 1
2A
= 2/A2 (4.28)
jz 802
9=o jz (A-jO)9 a=0
58 Chap. 4 Densities and Characteristic Functions
f I g(9) 12 dB = 1 (4.34)
This is a very basic result, and will be important to our considerations since the g's,
will turn out to be "signals".
We show the sufficiency here because it produces a result that is revealing. As-
suming that M(9) is a proper characteristic function, the density is given by
Sec. 4 Two Dimensional Densities 59
a-jax (4.35)
P(x) = 2_ f M(O) dO =
2- ff g'(B')g(9' + 9) a-jex d8' dO
which shows that P(x) is a proper density since it is positive. By Parceval's theorem
we know that P(x) is normalized to one if g is normalized to one as per Eq. (4.34).
Note that while there is a one-to-one relationship of the density to the character-
istic function, that is not the case with the function g. Many g's produce the same
characteristic function and hence the same density.
Again we normalize to one so that, for example, if P(x, y) is the number of people of
a certain weight and height, then when normalized to one it becomes the fraction
of people of that weight and height.
Marginals. If we have a joint density we may want the density of just one of the
variables irrespective of the value of the other variable. This is achieved by inte-
grating out the other variable,
These densities are called marginal densities or marginals. The origins of the word
"marginal" came about as follows. When the basic concepts of probability and
statistics were being developed, joint densities were written on a piece of paper
in columns and rows, in the fashion of the modem spreadsheet. For example, if
we wanted to make a table indicating how many individuals there are of a certain
height and weight (the joint density), the weight axis would be on top of the page
and the height axis would run down the page as in a typical spreadsheet. At the
60 Chap. 4 Densities and Characteristic Functions
intersection of each row and column one would write the number of people of that
weight and height. Now suppose one wanted the number of people of a certain
weight irrespective of height (the weight marginal) and the number of people of a
certain height, irrespective of weight (the height marginal). Then one adds up the
columns and rows. A natural place to put the sums is in the margins of the paper,
that is on the right margin and the bottom of the page - hence the terminology
marginals.
M(9, r) _ [ [I-
00
_ (7e)n(.7 r)m
00
nim! (xnym) (4.42)
n=0 m=0
and therefore
1 an+m
(4.43)
(xnym) jnjm 80"(97- M(9, T) 9,T = 0
Relation Between the Joint Characteristic Function and the Characteristic Func-
tion of the Marginals. The characteristic function of the marginal for x is
with Irl < 1. The characteristic function, calculated by way of Eq. (4.40), is
M(O, r) = exp [ jaB + jbr - (a'9' + 2rixvyOr + a 2T2) ] (4.47)
2
The characteristic function of the marginal in x is therefore
M(9) = M(9, 0) = exp [jaO - o 9 2 / 2 ] (4.48)
This is the same form as the characteristic function of a one dimensional Gaussian, Eq.
(4.22), and hence the marginal is Gaussian with mean a and standard deviation a..
This is recognized to be the characteristic function of a Gaussian, Eq. (4.22), with mean
a + b and standard deviation ay + 2ro- o + Qy.
62 Chap. 4 Densities and Characteristic Functions
For this situation the characteristic function and the joint moments are also fac-
torable,
M(9,T) = M=(9)MY(T) (x"ym) = (x')(ym) (4.53)
However, one should be careful because sometimes a density may appear factorable
without being independent. For independence the factors must be the marginals.
The examples below illustrate this point.
Consider first
4
74xy if0< x,y <a
(4.54)
0 otherwise
The marginals are
4 2 4
fQ 2
P(x) = IQ xy dy = Q2 x P (3l) = Q4 xy = Q2 y (4.55)
a4
Thus P(x, y) = P(x)P(y) and hence the distribution is an independent one. Now
consider
4 0 C X, y
T4
xy if x2 + y2 < a2 ;
P(x, y) = (4.56)
0 otherwise
We have
2-y2
P(x) = a4
xJ
I0
V,.2-y2
y dy = Q4 x (a2 - x2) (4.57)
P(y) = 8
a4
yJ x dx = 4
a
y (a2 - y2) (4.58)
o
We see that even though the joint density is a product in some sense it is not inde-
pendent because the product is not the product of the marginals. In fact, this joint
distribution is not really factorable for all values of x and y. Examples like these have
historically been presented as a warning to use caution. In reality it is just a failure of
notation. If we rewrite the joint density as
4
P(x, y) = W,-x y E(a2 - x2 - y2) E(x) E(Y) (4.59)
where c(x) is the step function, then it is clear that P(x, y) is not factorable and the
issue never arises.
Sec. 5 Local Quantities 63
Covariance and Correlation. Suppose we want to know how strongly one variable
depends on the other. This is what a joint distribution tells us in precise terms.
However, if we want a gross measure that reflects in a simple way the strength of
the dependence, we can consider the expected value of the first mixed moment,
(x y). If the joint distribution is independent, the first mixed moment will equal
the product of the moments,
(x y) = (x) (y) [ x, y independent ] (4.60)
The dependency can be measured by the excess of the first mixed moment over
(x) ( y ). That is the covariance
The covariance is a number that can be positive or negative and of any magnitude.
To standardize the correlation coefficient is defined by
COVxy
r = (4.62)
QxQy
where Qx and ay are the standard deviations of x and y. The correlation coefficient
ranges from -1 to +1 for all densities and hence standardizes the strength of the
correlation between x and y. The variables x and y are said to be uncorrelated if the
correlation coefficient is zero, strongly correlated at r = + 1, and strongly oppositely
correlated when r = -1. However, one should be cautious in applying these ideas
because they do not necessarily reflect the intuitive notion as to whether the two
variables have something to do with each other. One must not confuse dependence
with correlation. In fact, we can have a distribution where it is dear that the two
variables are strongly dependent and yet the correlation coefficient is zero. The
correlation coefficient is a single number and we cannot expect too much from just
one number. Nonetheless, the correlation coefficient is often a good gross indicator
of the dependence of two variables. The simplest way to calculate the first mixed
moment is from the characteristic function
492M(9, T)
(x y) a00T
(4.63)
9,r = 0
if it is available.
Conditional Averages and Their Standard Deviation. Suppose we want the aver-
age of y for a given x, for example, the average weight of those people who are 6
ft. tall. That is called the conditional average and is denoted by (y),,. Since the
density of y for a given x is P(x I y), we have
More generally, the conditional average of any function (the local average) is
a2y1x
1
(y2)r-(y)2 (4.68)
This equation shows that the global average is the average of the conditional aver-
age.
We now seek the relationship between the global standard deviation and the
conditional standard deviation. The global standard deviation is the not average of
the local standard deviation, but rather
(Y). (4.71)
Q2 = f vy21x P(x) dx + J - (y ) )2 P(x) dx
To prove this we average er ,
Now subtract and add (y ) 2 to the right hand side of Eq. (4.73),
This solves the problem because from P(x) we obtain Mu(O) and then P(u) by
Eq. (4.77). We can actually carry out the procedure explicitly and obtain an interest-
ing form for P(u). Substitute Eq. (4.78) into Eq. (4.77), to obtain
or
P(u) = 6(u - f (x)) P(x) dx (4.80)
J
The physical interpretation is dear. The delta function picks up from P(x) only the
values for which u = f (x). This form avoids the calculation of the characteris-
tic function. Further simplification is possible. One of the properties of the delta
function is
6(g(x)) = 6(x - xi) (4.81)
Ig'(xi)'
where xi's are the solutions to g(Ti) = 0, that is, the zeros of g. In our case we have
g(x) = f (x) - u and since the derivative of g is the derivative of f we have
(4.82)
i)
where now the xi values run over the solutions of f (xi) = u. Therefore,
or
Suppose there is only one solution to f (xi) = u, which we might as well call x.
Then
_ P(x) (4.86)
P(u) I P(x) x = f (u)
where x = f - (u) signifies that we must solve for x in terms of u by solving f (x) =
u. Equivalently, we note that for this case P(u) du = P(x) dx, which is the easiest
way to do it.
Sec. 7 Distribution of a New Variable 67
Since for the range considered the function is single valued, we have f' (x) = -A sin x
and therefore
P(u) = 1 1 1 1
if -A<u<A (4.89)
7r A sin x j x=arccoe(u/A) 7r .J42 - u
xl = U, x2 = -u (4.90)
Suppose we wish to scale a variable and also translate it. We take u = ax + b. The
characteristic function of u is
P(u) = I f e-jauMU(0) dO = a P. (u b) a1
(4.93)
T7r a
Alternatively, we can use Eq. (4.86). Solving for the zero we have x = (u - b)/a and
hence
Two Dimensional to One. Suppose we have a two dimensional density P(x, y) and
wish to find the density of a single variable which is a function of x and y,
u = f(x,y) (4.95)
P(u) = (4.97)
= JP(xtL_x)dx (4.100)
Two Dimensional to Two Dimensional. In the most general case we start with a
density of two variables and wish to find the density of two new variables function-
ally related by
u = f(x,y) ; v = g(x,y) (4.101)
Using the characteristic function approach we have, for the characteristic function
for the variable u and v,
eJef (x,v)+j'rg(x,v) P(x, y) d x dy (4.102)
M., (e, T) _ Jf
The density for u, v is then
5.1 INTRODUCTION
The aim of this chapter is to show, by way of examples, the need for a combined
time-frequency representation and to clarify why time analysis and frequency anal-
ysis by themselves do not fully describe the nature of signals. We also describe some
of the physical reasons why spectra change in time.
Suppose we have the individual densities of height and weight of a particu-
lar type of animal. These individual densities tell us everything we want to know
about the distribution of height and the distribution of weight. From these distri-
butions can one determine how height and weight are related? Can one determine
whether the tall individuals are the heavy ones? The answer is no. Can one de-
termine whether an individual with weight of 250 lb. is exceptional? The answer
is yes. But can one determine whether he is exceptional for his height? The an-
swer is no. The individual densities of height and weight are not a full description
of the situation because from these densities we cannot ascertain how height and
weight are related. What is needed is the joint density of height and weight. In
the same way, the time energy density and the frequency energy density are not
sufficient to describe the physical situation because they do not fully describe what
is happening. In particular, from the spectrum we know which frequencies were
present in the signal but we do not know when those frequencies existed, hence the
need to describe how the spectral content is changing in time and to develop the
physical and mathematical ideas to understand what a time-varying spectrum is.
We wish to devise a distribution that represents the energy or intensity of a signal
simultaneously in time and frequency.
Time varying spectra are common in ordinary life. During a sunset, the fre-
70
Sec. 2 Simple Analytic Examples 71
quency composition of the light changes quickly and dramatically. In saying that
the sky is getting redder we are conveying a time-frequency description because we
are describing how the frequencies are changing in time. The pitch, which is the
common word for frequency of human speech, changes as we speak and produces
the richness of language. Similarly the pitch of animal sounds changes during vo-
calization. Standard musical notation is a time-frequency representation since it
shows the player what notes, or frequency, should be played as time progresses.
In this chapter we give a number of simple analytic and real examples of time
varying spectra so that we may become familiar with the need for a time-frequency
description, the variety of applications, and the language involved. We leave aside
the methods used to construct the joint time-frequency representations. That will
be the subject of the subsequent chapters. The reader should think of these repre-
sentations as no different from a joint density of height and weight of a population
that gives the relative concentration of people at particular heights and weights.
Similarly we should think of a time-frequency distribution as telling us the inten-
sity or energy concentration at particular times and frequencies.
We first take a real example to illustrate the basic idea before embarking on a
more systematic discussion. Using a whale sound, Fig. 5.1 shows three plots. Run-
ning up the page is the sound (the air pressure) as a function of time. By examining
it visually we cannot tell much, although we can clearly tell how the intensity or
loudness varies with time. Below the main figure is the energy density spectrum,
that is, the absolute square of the Fourier transform. It indicates which frequencies
existed and what their relative strengths were. For this sound the spectrum tells us
that the frequencies ranged from about 175 to about 325 cycles per second. This in-
formation is interesting and important, but does not fully describe what happened,
because from the spectrum we cannot know when these frequencies existed. For
example, we cannot know just by looking at the spectrum when the 300 Hz sound
was made, or whether it was made for the total duration of the sound or just at
certain times. The main figure is a time versus frequency plot, that is, a joint time-
frequency distribution. From it we can determine the frequencies and their relative
intensities as time progresses. It allows us to understand what is going on: At the
start the frequency was about 175 Hz and increased more or less linearly to about
325 Hz in about half a second, stayed there for about a tenth of a second, and so
forth. In answer to the question of when the 300 Hz sound occurred, we can now
see the answer. It occurred twice, at 0.6 and 1.3 seconds.
The difference between the spectrum and a joint time-frequency representation
is that the spectrum allows us to determine which frequencies existed, but a com-
bined time-frequency analysis allows us to determine which frequencies existed at
a particular time.
1.4 H
0.4 H
0.2
Fig. 5.1 A time-frequency plot of a sound made by a Bowhead whale. The wave
form is on the left plot with time increasing upwards. The energy density spec-
trum is below the main figure and indicates which frequencies existed for the dura-
tion. The main figure is a joint time-frequency plot and shows how the frequencies
change with time. (From the work of C. Rosenthal and L. Cohen.)
Finite Duration Sine Waves. In Fig. 5.2 we show examples of signals composed of
finite duration sine waves at three frequencies. All three examples have basically
the same spectrum, which is reasonable since for all cases we had three frequencies.
The cases are different in regard to when the frequencies existed and this can be
easily ascertained from the time-frequency plots.
s(t) = 1
e'At2/2+jwot 0<t<T (5.1)
vT
1 2 3
(a) FREQUENCY
I A A A
where Jn is the nth-order Bessel function. Fig. 5.3(b) plots the signal energy density
spectrum and the time-frequency plot. There is a considerable difference between
what we learn from the two.
Other Examples. The cases of a cubic phase and a log phase are shown in Figs. 5.3
(c) and (d).
0 10
(d)
so
0 4
W
E 10
Fig. 5.3 The time-frequency plots for various signals. In (a) we have a chirp, s(t) =
eipt2/2+,wot.
The instantaneous frequency, m, = .o +$t, is increasing linearly. The
energy density spectrum is more or less flat, indicating that each frequency existed,
but gives no indication of when it existed. The time-frequency plot does. In (b)
we have a sinusoidal modulation, e'Ot2/2+,m sin Wmt+ikJot, and the energy spectrum
shows a concentration at certain frequencies but gives no indication of the times.
The reason that there is concentration at those frequencies is that those frequencies
last longer, as can be clearly seen from the time-frequency plot In (c) we have a
i,yt3/3+Ot2/2+iwot.
signal with a cubic phase, s(t) = In (d) we have a hyperbolic
signal,s(t) = e1 In(t-to) which gives an instantaneous frequency ofwt = 1/(t-to),
which is a hyperbola in the time-frequency plane.
Mn
1I`
Human speech. The analysis of human speech was the main reason for the practical
development in the 1940s of time-frequency analysis. The main method was and
still is the short-time Fourier transform, which we will study in Chapter 7. This
approach gave dramatic new understanding of the production and classification of
speech. An example is shown in Fig. 5.5.
08
06
04
02
0
0 50 100 150 200
0
Bowhead Whale
Propeller Sounds. When a propeller or fan turns it pushes the air at a repetitive
rate depending on the rotation. If the propeller has, for example, three symmetrical
blades, then the frequency of repetition will be three times the number of revolu-
tions per unit time. For a constant revolution rate this will show up as an acoustic
pressure wave of constant frequency. However, if the ship decreases or increases
its speed by changing the revolution rate this will result in changing frequencies.
Inelastic Acoustic Scattering. When a rigid object is hit with a sound wave, the
wave gets reflected, producing an echo. The characteristics of the reflected wave
depend on the shape of the object, its surface properties, the medium of propa-
gation, and location. This reflection is called specular reflection. However, if the
object is deformable, then the initial wave that hits the object produces waves in
the object no differently than if the object had been hit with a hammer. The waves
travel within the object and get reradiated. Hence for an elastic object hit by a wave
not only do we get the usual echo, but we get additional waves due to the elastic
response. Fig. 5.7 gives an example of a toy submarine which is a thin metallic shell
with internal supporting structures. The submerged submarine was irradiated with
an acoustic wave (sonar). The first part of the returned signal is the specular reflec-
tions. The rest of the signal is due to the reradiated acoustic elastic response. The
advantage of a time-frequency plot is that we can immediately see at what times
the reradiation occurred and what frequencies were radiated out.
Windshield Wipers. The sounds that windshield wipers make are often annoying.
Fig. 5.8 shows a time-frequency plot of a good blade and an annoying one. In the
annoying case a web-like structure is seen at higher frequencies. The frequency
band appearing in both is due to the motor sound.
Sec. 3 Real Signals 77
Fig. 5.8 Time frequency plots of windshield blades of two different models of cars.
The first (a) is judged to be annoying, the second (b) is not. (Courtesy of J. Feng,
Ford Motor Co.)
Car Door Slams. The slamming of a car door has historically been a major concern
for many reasons, Madison Avenue notwithstanding. We have all experienced the
rich sounding door with its solid thump and the tinny door that does not engen-
der room for confidence in the rest of the construction. The sound produced by a
slamming car door is quite involved because there are a number of diverse events
occurring including the coupling with the latch, the hinging, and the seals around
the frame. Fig. 5.9 shows a time-frequency plot for a good sounding car door (a) and
a tinny one (b). Note that in the poor sounding door there is considerable acoustic
energy at a wide spread in frequencies immediately after the contact of door with
latch. It has been found that this acoustic energy, which lasts for about a twentieth
of a second, is the main reason people say a door sounds tinny and hence poor. The
main frequency concentration is the vibration of the automobile as a whole caused
by the slamming of the door.
(a) (b)
-* t
Fig. 5.9 Time-frequency plots of two car door slams. The one in (a) is characterized
by a tinny sound and the one in (b) by a rich sound. The tinny sound is caused by
the generation of a wide spectrum of frequencies in the initial contact of the door
with the latch. (Courtesy of J. Feng, Ford Motor Co.)
78 Chap. 5 The Need for Time-Frequency Analysis
Fault Analysis. It is important to have methods for the earliest possible detection
of when a machine is starting to go bad. This is particularly important for machines
such as in airplanes and ships whose failure may lead to critical situations. It is also
critical in manufacturing since a bad manufacturing machine may produce defec-
tive products. We illustrate such situations in Figs. 5.10 and 5.11.
(a)
71
(a)
Fig. 5.11 Drilling of certain metal
parts often requires the resulting
hole to be of high precision. There-
fore it is of importance to have an ef-
(b) fective means for the determination
as to when a drill is beginning to get
dull. In (a) we a have sharp drill, in
(b) the drill is somewhat dull and in
(c) the drill is very dull. The data is
obtained from an accelerometer on
I!: i .c,19A
the spindle. (Courtesy of P Lough-
(c) lin, J. Pitton, L. Atlas and G. Bernard.
..... ., nil-I w.. ..4J. Data courtesy of Boeing Commerical
f
Airplane Group.)
Sec. 3 Real Signals 79
Trumpet. In Fig. 5.12 we show a time-frequency plot of the first half second of a
trumpeter initiating a note. The ridges in the frequency direction are the harmonics
of the fundamental. To achieve the steady state desired the trumpeter has to adjust
in a split second. The frequency of each harmonic gets adjusted in the same way,
that is, the percentage change is the same. While the absolute change is insignificant
for the lower frequencies, it is significant at the higher frequencies.
Bottlenose Dolphin. Fig. 5.13 shows the time-frequency plot of a dolphin sound.
10
Heart Sound. While the metaphysical importance of the heart has been recognized
since ancient times, the notion that the heart is responsible for the heart sound was
not accepted until the 1700s. There was considerable debate as to whether the heart
produces a sound at all. Indeed, the invention of signal analysis can be traced to
Hook' :"I have been able to hear very plainly the beating of a Man's heart... Who
knows, I say, but that it may be possible to discover the Motions of the Internal Parts
of Bodies... by the sound they make, that one may discover the Works performed
in the several Offices and Shops of a Man's Body, and thereby discover what In-
struments or Engine is out of order". 16181 Fig. 12.4 shows a time-frequency plot of a
heart sound.
'in textbooks Robert Hook is remembered for Hook's law. However, he was one of the greatest
scientists of his century Hook discovered that plants were made up of cells, the red spot of Jupiter, the
wave nature of light, and Boyle's law of gases (Boyle published this law in a book on the subject and
gave full credit to Hook, but Boyle's name stuck), among many other discoveries. He was about ten
years older than Newton and their lives were intertwined both positively and negatively.
Sec. 4 Why Spectra Change 81
Time-Frequency Distributions:
Fundamental Ideas
6.1 INTRODUCTION
The basic objective of time-frequency analysis is to devise a function that will de-
scribe the energy density of a signal simultaneously in time and frequency, and that
can be used and manipulated in the same manner as any density. If we had such a
distribution we could ask for the fraction of the energy in a certain frequency and
time range, we could calculate the density of frequency at a particular time, we
could calculate the global and local moments of the distribution such as the mean
conditional frequency and its local spread, and so on. We now begin our study of
how to construct such distributions and in this chapter we describe the main ideas.
To crystalize our aim we recall that the instantaneous power or intensity in time is
82
Sec. 1 Introduction 83
Do there exist joint time-frequency distributions that satisfy our intuitive ideas
of a time-varying spectrum? How can they be constructed? Can they be interpreted
as true densities? Do they represent the correlations between time and frequency?
What reasonable conditions can be imposed to obtain such densities? The hope is
that they do exist, but if they don't in the full sense of true densities, what is the
best we can do? Are there inherent limitations to such a development? This is the
scope of time-frequency analysis.
l2
f P(t, w) dt = I S(w) (6.2)
Total Energy. The total energy of the distribution should be the total energy of the
signal
Note that if the joint density satisfies the marginals, it automatically satisfies the
total energy requirement, but the converse is not true. It is possible that a joint
density can satisfy the total energy requirement without satisfying the marginals.
The spectrogram that we study in the next section is one such example. The total
energy requirement is a weak one and that is why many distributions that do not
satisfy it may nonetheless give a good representation of the time-frequency struc-
ture.
However, many of the distributions that we will be studying are not proper. By
proper we mean a well-defined manifestly positive density function. Hence, for a
particular distribution, the characteristic function may not satisfy all the standard
attributes of characteristic functions that we discussed in Chapter 4.
84 Chap. 6 Time-Frequency Distributions: Fundamental Ideas
and of course the answers should be meaningful and reasonable. If the marginals
are satisfied then we are guaranteed that averages of the form
fi(t)l s(t) I2
dt + f 92 (w) I S(w)
I2
dw (6.7)
will be correctly calculated since the calculation requires only the satisfaction of the
marginals.
But what about averages that are mixed, that is, average values of arbitrary time-
frequency functions? An example is the covariance which involves the calculation
of ( t w ). For mixed averages we do not know what we want and can only be guided
by intuition and by the plausibility of possible guesses. That is not surprising. Math-
ematics cannot answer what is essentially an issue of science. Consider, for example,
the question of height and weight of an unknown species and suppose we knew the
distribution of height and weight, that is, the marginals. Can we know how height
and weight are related, how the correlation coefficient behaves? No. Similarly from
knowing the time-frequency marginals we cannot know how time and frequency
should be related and mathematics cannot help since mathematics allows all pos-
sibilities. In Chapter 1 we discussed the covariance of a signal and showed how
that is related to the first mixed moment (t w ). We argued that a plausible guess to
(t w) is the average of t gyp' (t) and we showed that this quantity behaves reasonably
well in the sense that it meets our intuition for the cases considered. Therefore, a
plausible requirement of a joint distribution is
Notice that we have used P(t) and P(w) rather than I s(t) 12 and S(w) 12 to allow
for the possibility that the marginals are not satisfied, as will be the case with the
spectrogram discussed in the next chapter.
The conditional average value of a function at a given time or frequency is
(g(t ) ). = P(w)
f 9 (t) P(t , w ) dt (6 . 12)
In Chapter 2 we saw that a good candidate for the average frequency at a given time
is the derivative of the phase, the instantaneous frequency. Similarly, we argued
that the average time for a given frequency should be the derivative of the spectral
phase. If we accept these results, then a joint time-frequency density should satisfy
(t). =
P(W)
ft P(t, w) dt = -ii'(w) (6.14)
From this point of view instantaneous frequency is an average, the average fre-
quency at a particular time. We should also be able to define the spread of fre-
quencies at a given time and the spread of time for a given frequency, that is, the
conditional standard deviations. These are
2 1
it ( w - (w) ) 2 P ( t , w) dw (6. 15 )
P(t)
1
2
at1w (t - (t) ) 2 P(t, w) dt (6.16)
P(w)
What should these quantities be? These quantities will play a crucial role in the
development of the basic theory, so we hope they turn out to be sensible. For some
signals we know what the answer should be. For example, for a pure sinusoid we
should get v,,It = 0, since a pure sinusoid has one definite frequency for all time.
if S(w) --+ S(w - wo) then P(t, w) --+ P(t, w - wo) (6.18)
Both of these cases can be handled together. If s(t) is the signal, then a sig-
nal that is translated in time by to and translated in frequency by wo is given by
e&"Ots(t - to). Accordingly, we expect the distribution to be shifted in time and
frequency in the same way,
S. (W) =Vfa_
I S(w/a) if ssc(t) = f s(at) (6.20)
We see that if the signal is compressed then the spectrum is expanded, and con-
versely. If we want these relations to hold for the joint distribution, then we must
have
P. (t, w) = P(at, w/a) (6.21)
The scaled distribution satisfies the marginals of the scaled signal
J2
Psc(t, w) dt = J Ssc(w) = a 1 S(w/a) 12 (6.23)
J
the spectrum is zero outside a frequency band, then the distribution should also be
zero outside the band. In such a case we say that the distribution has weak finite
spectral support. We can express these requirements mathematically as
P(t, w) = 0 fort outside (tl, t2) if s(t) is zero outside (tl, t2) (6.24)
P(t, w) = 0 for w outside (wl, w2) if S(w) is zero outside (wl, w2) (6.25)
Now suppose we have a signal that stops for a half hour and then starts again.
We would expect the distribution to be zero for that half hour. Similarly, if we have
a gap in the spectrum, then we expect the distribution to be zero in that gap. If a
distribution satisfies these requirements, namely that it is zero whenever the signal
is zero or is zero whenever the spectrum is zero, then we say the distribution has
strong finite support:
and the third is that s(t) and S(w) are Fourier transform pairs. From a joint distri-
bution the standard deviations are obtained by
When do we get the correct uncertainty principle? When the standard deviations
calculated using the joint distribution give the same answer as when calculated by
Eqs. (6.28) - (6.29). This will be the case when the marginals are correctly given,
Therefore, any joint distribution that yields the correct marginals will yield, and is
totally consistent with, the uncertainty principle.
_
P (x) = L
exp
r
_ (x 277 /1 )
z
J
(6 . 34)
z
-
P(y) = exp 1 2 77(12 + k2)/l
(6 . 5)
27r77(12 + k2 )/1
where l and k are any two positive numbers at our disposal. The universal constant
rl is fixed and not at our disposal.
These two marginals are perfectly good densities. But no matter what we do or
try to do we have the uncertainty principle, ayQx > rl. That is the case since
'The more general case is where is a functional of the signal. In the case of time-frequency it is a
constant equal to half. For time and scale, for example, it is functional of the signal. See Section 15.5. To
keep things simple we taken as a constant here but our discussions apply generally
Sec. 8 The Uncertainty Principle and Joint Distributions 89
Example 1. Consider the following perfectly well behaved joint distribution which
satisfies the marginal and hence the uncertainty principle,
P(x, y) =
1
exp - (x - a)2 _ (y - b)2
27r77 .V/1 + 12/k2 [ 2r7/1 21](12 + k2)/l ] P(x) P(y) (6.38)
=
This distribution is the independent joint distribution. Even though the marginals
are related, the variables are not correlated. For any value of x the density of y is
the same.
Example 2. Another joint density, which satisfies the same marginals and has cor-
relations, is
1
y) (6.39)
P(x ' = 2ir771_12/k2_,/"1
+ - r2
1 (x - a)2 r(x - a)(y - b) (y - b)2
exp -2 +
- 2(1 - r2) rl/l 77
1 + l2/k2 77(l2
+ k2)/l ( 6 . 40)
where r is any number between -1 and 1. For this case there is correlation be-
tween the variables. The correlation is positive or negative, depending on the sign
of r. There is a rich variety of possibilities, but the marginals are always satisfied
90 Chap. 6 Time-Frequency Distributions: Fundamental Ideas
Fig. 6.1 The uncertainty principle depends only on the marginals. It constrains the
type of joint distribution that is possible. A joint distribution cannot be narrow in
both variables, but this still leaves an infinite number of possibilities. The marginals
give no indication as to whether, or how, two variables may be correlated.
and so is the uncertainty principle. In particular, when one marginal is broad, the
other is narrow. A proper joint distribution has no difficulty accommodating such
marginals and although we have just demonstrated two of them, there are an infi-
nite number of them. 11771 Lest the reader think this is an artificial example, we point
out that it is precisely the Wigner distribution for a chirp with a Gaussian envelope.
Also, constructing other joint distributions is very easy and we do so in Section 14.2.
In Fig. 6.1 we plot the distribution for two different sets of marginals and dif-
ferent values of l and k. For a given set of marginals, one can always find an infi-
nite number of joint distributions with various correlations between the variables.
From the marginals, nothing can be concluded about the correlation between the
variables.
there is a relationship between the global standard deviation and the conditional
standard deviation and the conditional average. In particular, writing Eq. (4.3) for
both x and y we have
Note that for each variable there are two ways to control the standard deviation and
hence when we multiply ci by o we will have four terms, the sum of which must
be greater than r72. We can satisfy the uncertainty principle by making any one of
those terms greater than ,772 and have freedom with the rest of them. Or, in general,
we can choose any value for each term as long as the sum adds up to greater than
77 2. These four terms are a gross characterization of a joint distribution. Because
they are not individually constrained we can have a wide variety of qualitatively
different joint distributions and still satisfy the uncertainty principle.
Signal Analysis. The case of signal analysis is no different, although historically the
uncertainty principle was viewed as presenting unsurmountable difficulty for the
construction of joint distributions. Quite the contrary, it is trivially easy to construct
joint distributions consistent with the uncertainty principle, as the above consider-
ations have shown. All we have to do is satisfy the marginals. In fact, the real
problem is that there are an infinite number of such joint distributions and we do
not fully understand what other considerations besides the uncertainty principle
are required to construct a sensible and comprehensive theory. Satisfying the un-
certainty principle is easy; the rest is challenging.
The current state of affairs is that we do not have a complete theory. Nonethe-
less, the ideas and techniques that have been developed thus far are powerful, give
us considerable insight into the nature of signals, meet our intuition to a very large
extent, and have been applied with immense success to practical problems. This is
demonstrated in the subsequent chapters. Often some of the predicted results are
clearly not plausible and that is what makes the subject challenging and fascinat-
ing. Because we do not have a complete and comprehensive theory, it is important
to understand what is known with certainty, what is speculation, and what hidden
assumptions go into any particular proof.
Since we have not yet studied any distributions it would not be appropriate at
this juncture to discuss the mathematical or physical issues in any great detail. To
appreciate these issues fully, we should first see how they arise. We do mention one
idea that has a checkered history. Because many distributions have been proposed
over the last fifty years, something that should be considered in a positive light,
there have been attempts to prove that a particular one is best. This has been done
by listing a set of desirable conditions and trying to prove that only one distribution,
usually the writer's favorite, fits them. Typically, however, the list presented is not
complete with the obvious requirements, because the author knows the added de-
sirable properties would not be satisfied by the distribution he or she is advocating.
Also these lists very often contain conditions that are taken out of thin air and are
obviously put in to force a particular result. Moreover, there are usually hidden as-
sumptions in the proofs that are glossed over or hidden in the methodologies of the
proofs. However, there is one benefit to these impossibility and uniqueness proofs.
The searching of the errors and hidden assumptions sometimes leads to a reexam-
ination of the basic ideas. There are time-varying spectra in nature and therefore
their description and properties can be described and understood. Someone will
eventually come up with the right theory and when that happens we will not have
to prove it mathematically correct; it will be obviously correct.
Terminology. For the type of distributions we will be studying, a rich variety of ter-
minology has arisen. We first address the general usage of density and distribution.
When we say that a function represents the number of things per unit something,
that is called a density and hence density function. In certain fields density func-
tions are called distributions because they indicate how things are distributed. For
example, one says the "Maxwell distribution of velocities." As discussed in Chapter
4, we shall use the words density and distribution interchangeably. These types
of distributions first arose in quantum mechanics where the term "probability den-
sity" or "distribution" is properly applied, since quantum mechanics is inherently
probabilistic. For deterministic signals where no probabilistic considerations enter,
the reader should think of distributions as "intensities" or "densities" in the com-
mon usage of the words, or simply as how the energy is "distributed" over the
time-frequency plane. As we will see, many of the known distributions may be-
come negative or even complex. Hence they are sometimes called quasi or pseudo
distributions. Also, a joint time-frequency distribution is, of course, dependent on
the signal and can be said to represent the signal in time and frequency; hence the
phrase "time-frequency representation."
Chapter 7
7.1 INTRODUCTION
The short-time Fourier transform is the most widely used method for studying non-
stationary signals. The concept behind it is simple and powerful. Suppose we listen
to a piece of music that lasts an hour where in the beginning there are violins and
at the end drums. If we Fourier analyze the whole hour, the energy spectrum will
show peaks at the frequencies corresponding to the violins and drums. That will tell
us that there were violins and drums but will not give us any indication of when the
violins and drums were played. The most straightforward thing to do is to break
up the hour into five minute segments and Fourier analyze each interval. Upon
examining the spectrum of each segment we will see in which five minute intervals
the violins and drums occurred. If we want to localize even better, we break up the
hour into one minute segments or even smaller time intervals and Fourier analyze
each segment. That is the basic idea of the short-time Fourier transform: break
up the signal into small time segments and Fourier analyze each time segment to
ascertain the frequencies that existed in that segment. The totality of such spectra
indicates how the spectrum is varying in time.
Can this process be continued to achieve finer and finer time localization? Can
we make the time intervals as short as we want? The answer is no, because after
a certain narrowing the answers we get for the spectrum become meaningless and
show no relation to the spectrum of the original signal. The reason is that we have
taken a perfectly good signal and broken it up into short duration signals. But short
duration signals have inherently large bandwidths, and the spectra of such short
duration signals have very little to do with the properties of the original signal. This
should be attributed not to any fundamental limitation, but rather to a limitation
of the technique which makes short duration signals for the purpose of estimating
93
94 Chap. 7 The Short-Time Fourier Transform
the spectrum. Sometimes this technique works well and sometimes it does not. It is
not the uncertainty principle as applied to the signal that is the limiting factor; it is
the uncertainty principle as applied to the small time intervals that we have created
for the purpose of analysis. The distinction between the uncertainty principle for
the small time intervals created for analysis and the uncertainty principle for the
original signal should be clearly kept in mind and the two should not be confused.
We should always keep in mind that in the short-time Fourier transform the
properties of the signal are scrambled with the properties of the window function,
the window function being the means of chopping up the signal. Unscrambling is
required for proper interpretation and estimation of the original signal.
The above difficulties notwithstanding, the short-time Fourier transform method
is ideal in many respects. It is well defined, based on reasonable physical principles,
and for many signals and situations it gives an excellent time-frequency structure
consistent with our intuition. However, for certain situations it may not be the best
method available in the sense that it does not always give us the clearest possible
picture of what is going on. Thus other methods have been developed, which are
discussed in subsequent chapters.
The modified signal is a function of two times, the fixed time we are interested in,
t, and the running time, T. The window function is chosen to leave the signal more
or less unaltered around the time t but to suppress the signal for times distant from
the time of interest. That is,
st(T) -
s(r) for 7- near t
0 for T far away from t
The term "window" comes from the idea that we are seeking to look at only a small
piece of the signal as when we look out of a real window and see only a relatively
small portion of the scenery. In this case we want to see only a small portion.
Since the modified signal emphasizes the signal around the time t, the Fourier
transform will reflect the distribution of frequency around that time,
7
St(w) = st(T) dT
1
a-CWT s(T) h(T - t) d7- (7.4)
727
Sec. 2 The Short-Time Fourier Transform and Spectrogram 95
This shows that the spectrogram can be used to study the behavior of time proper-
ties at a particular frequency. This is done by choosing an H(w) that is narrow, or
equivalently by taking an h(t) that is broad.
Narrowband and Wideband Spectrogram. As just discussed, if the time window
is of short duration the frequency window, H(w), is broad; in that case the spec-
96 Chap. 7 The Short-Time Fourier Transform
is the ambiguity function of the signal, and Ah is the ambiguity function of the
window defined in the identical manner, except that we use h(t) instead of s(t).
Note that A(-O, T) = A (0, -r), a relation we will use later.
Notation. The results we will obtain are revealing when expressed in terms of the
phases and amplitudes of the signal and window and their transforms. The nota-
tion we use is
s(t) = A(t) Ow(t) h(t) = Ah(t) eawn(t) (7.13)
S(w) = B(w) H(w) = BH(w) erLX(W) (7.14)
and similar notation will be used for other quantities. When it is clear from the
context which density is being used the superscript notation will be omitted.
Therefore, we see that if the energy of thelwindow is taken to be one, then the
energy of the spectrogram is equal to the energy of the signal.
P(t) = f I St(w)
I2
dw (7.21)
=
2 -
f s(r) h(T - t)9*(r') h*(r' - t)
e-jc(T_T')drdr'
dw (7.22)
fs(r)2Ih(r_t)I2dr (7.24)
fA2(r)A2h(r_t)dr (7.25)
As can be seen from these equations, the marginals of the spectrogram generally do
not satisfy the correct marginals, namely I s(t) I2 and I S(w) 12,
The reason is that the spectrogram scrambles the energy distributions of the win-
dow with those of the signal. This introduces effects unrelated to the properties of
the original signal.
Notice that the time marginal of the spectrogram depends only on the magni-
tude of the signal and window and not on their phases. Similarly, the frequency
marginal depends only on the amplitudes of the Fourier transforms.
Averages of Time and Frequency Functions. Since the marginals are not satisfied,
averages of time and frequency functions will never be correctly given,
98 Chap. 7 The Short-Time Fourier Transform
Finite Support. Recall from our discussion in Chapter 6 that for a finite duration
signal we expect the distribution to be zero before the signal starts and after it ends.
This property was called the finite support property. Let us see whether the spectro-
gram satisfies this property. Suppose one chooses a time t before the signal starts.
Will the spectrogram be zero for that time? Generally, no, because the modified
signal as a function of t will not necessarily be zero since the window may pick up
some of the signal. That is, even though s(t) may be zero for a time t, s(r) h(r - t)
may not be zero for that time. This will always be the case for windows that are not
time limited. But even if a window is time limited we will still have this effect for
time values that are close to the beginning or end of the signal. Similar considera-
tions apply to the frequency domain. Therefore the spectrogram does not possess
the finite support property in either time or frequency.
One Window or Many? We have just seen that one window, in general, cannot
give good time and frequency localization. That should not cause any problem of
principle as long as we look at the spectrogram as a tool at our disposal that has
many options including the choice of window. There is no reason why we cannot
change the window depending on what we want to study. That can sometimes be
done effectively, but not always. Sometimes a compromise window does very well.
One of the advantages of other distributions that we will be studying is that both
time and frequency localization can be done concurrently.
Entanglement and Symmetry Between Window and Signal. The results obtained
using the spectrogram generally do not give results regarding the signal solely, be-
cause the short-time Fourier transform entangles the signal and window. Therefore
we must be cautious in interpreting the results and we must attempt to disentan-
Sec. 4 Global Quantities 99
gle the window. That is not always easy. In fact, because of the basic symmetry
in the definition of the short-time Fourier transform between the window and sig-
nal, we have to be careful that we are not using the signal to study the window.
The mathematics makes no distinction. The distinction must come only from a ju-
dicious choice of window, which is totally under our control. The basic symmetry
between the window and signal does have a mathematical advantage in that results
obtained should show a symmetry and therefore act as a mathematical check.
If the window is chosen so that its mean time and frequency are zero, which can be
done by choosing a window symmetrical in time and whose spectrum is symmetri-
cal in frequency, then the mean time and frequency of the spectrogram will be that
of the signal.
The second conditional moments are calculated to be
By combining these with Eqs. (7.32) we find that the duration and bandwidths are
related by
Tsp=Ta +Th ; B2P=B8 +Bh (7.35)
which indicates how the duration of the windowed signal is related to the durations
of the signal and window.
(t )(a)
((Ph
-(t)(h) + (t)(g) )(h) (7.37)
100 Chap. 7 The Short-Time Fourier Transform
Subtracting (t ) (SP) ( w) (SP), as given by Eqs. (7.32), from both sides we have that
the covariance of the spectrogram is
Cov " = (t w) (SP) - ( t ) (SP) ( w) (5P) = Cove) - CovtW) (7.38)
We know from Chapter 1 that the covariance of a real signal is zero and hence if we
take real windows, the covariance of the spectrogram will be the covariance of the
signal
CovtWP) = Covtw for real windows (7.39)
A fruitful way to look at the situation is that we are dealing with a signal in the
variable ,r whose amplitude is A(T) Ah(T - t) and whose phase is W(T) +Y'h(T - t).
However, st(T) is not normalized in the variable T and therefore we define, as in
Section 3.4, the normalized modified signal by
77t (.0
- S(T) h(r - t) - s(T) h(r - t)
(7.42)
f s(T) h(T - t) j2 dT P(t)
(w )t =
P(t)
f w St(w)
12
dw = f'i(r)'it(T)dT (7.44)
= P1
/
l A(T) Ah(T -
12
t)) dr
(T) dr
+ P1 IA'(,r)A'(,r-t)fw'(r)+wh(7--t)l'd7- (7.47)
P(t)
I (dT A(r) Ah(r - t) dT
+ 2P2I (t)
If A2(rl) A2(r2) Ah(r1 - t) Ah(7-2
- t)
Also, we consider real windows so as not to introduce phase due to the window. In
this limit, Eq. (7.45) goes as
f A2(7-)Ah(r-t)cp'(T)dr (7.52)
(w )t f A2(T) Ah (r - t) dr
f A2(r) 6(T - t) co'(r) dT A2(t) cp'(t)
(7.53)
f A2 (r) 6(T - t) dr A2(t)
or
(w)t-'cp'(t) (7.54)
That is, if the window is narrowed to get increasing time resolution, the limiting
value of the estimated instantaneous frequency is the derivative of the phase, which
is the instantaneous frequency. This a very pleasing and important result, but there
is a penalty to pay. As we narrow the window we broaden the standard deviation
and, therefore, as the estimate goes to the right answer it becomes harder to see and
determine. The reason for this can be seen by direct mathematical substitution of
the delta function in Eq. (7.49), which results in
2
Q
j" (w)
f(_B(w)Bh()_w))dw
ffB2(wi) B2(w2) Bh2.(w - wl) B2(w - W2)
+ 2P2(w)
x [ Y''(wl) - IP'(w2) - ?4 (W - wl) + Oh' (w - W2) ] 2 dwl dw2 (7.59)
7.8 EXAMPLES
Example 7.1: Sinusoid.
we have
St(w) = (a/7r) 1/4 a-)wt e-'(t-to)2/2 (7.65)
We have
o)2
St (W) = (a 11/4 e-)(W-w)t eXp [- (w 1/4 a-i&to e-a(t-to)2/2
2a 1 + (a/7r)
/l J (7.68)
and
This example illustrates one of the fundamental difficulties with the spectrogram. For
one window we cannot have high resolution in time and frequency. The broadness
of the self terms, the first two terms of Eq. (7.69), depends on the window size in an
inverse relation. If we try to make one narrow, then the other must be broad. That
will not be the case with other distributions. In Fig. 7.1 we plot the spectrogram for
different window sizes. The cosine term is an example of the so called cross terms. We
discuss these in greater detail later. Note that they essentially fall on the self terms in
the spectrogram.
PsP((-,w) = I St(w) I 2 = p
[_' JI
(7.71)
- (2
)w)2
P(w) t (7.72)
;7 0,21w
exp [- 2ortlw
20
10 10
0 0 0 0
(b)
20 20
10Q W010
00
Fig. 7.1 Spectrograms of a signal composed of a constant frequency plus an im-
pulse, s(t) = ej1ot + 6(t - 10). In (a) we use a short duration window, which gives a
good indication of when the impulse occurred but gives a broad localization for the
frequency. In (b) we use a long duration window, which gives the opposite effect.
In (c) a compromise window is used. For a comparison with the Wigner distribution
see Fig. 8.5.
P(t) - 7r(a a)
exp f - as t2 (7.73)
V
P(W) =
as/ _ as 21
ex p
p (7.74)
aa2 + a(a2 + 02) aa2 + a(a2 +,32) J
and
a 2
(w )t _ a + a pt + wo Uwlt = 2 (a + a) + 1 ap (7.75)
a
(t)om = aR
aa2 + a(a2 + T2
w (7.76)
2 1 (a + a)2 + Q2
t" 2 aa2 + a(a2 + N2) (7
The concentration of energy for a given time is along the estimated instantaneous fre-
quency, and for a given frequency it is along the estimated time delay. As the window
becomes narrow, that is, as a -+ oo, the estimate for the instantaneous frequency ap-
proaches pt + wo. However, in this limit the estimated group delay approaches zero,
which is a very bad estimate. In fact, this answer is just the average time for the sig-
nal. It is an understandable answer because in that limit we have a flat window in the
frequency domain. Conversely, if we want to focus in on time properties for a given
frequency we must take a broad window in the time domain, which is achieved by
taking a - oo. In that case (t )" _ 0/(a2 + p2), which is the correct answer (see Eq.
106 Chap. 7 The Short-Time Fourier Transform
(1.121)). On the other hand, at this limit the estimate of the instantaneous frequency
approaches wo, which is not the instantaneous frequency at all but just the average
frequency. Again, the reason is in that limit the window in the time domain is flat for
all time. This example shows that for the spectrogram one window cannot give good
localization in both time and frequency. In contrast, other distributions will allow us
to do that. We also note that for this case, the local bandwidth is a constant in time.
That would not be the case if a different window were chosen.
and
which indicates that the frequency oscillates sinusoidally around the straight line given
by (3t+wo. To calculate the estimated instantaneous frequency we use Eq. (7.45) to ob-
tain
St(w) =
()1/4
ack a-at'/2 (7.87)
where J are the Bessel functions. Substituting this into Eq. (7.87) and evaluating the
Gaussian integral we have
7.9 INVERSION
Can the signal be recovered from the spectrogram? Since the spectrogram is the
absolute square frequency of the short-time Fourier transform it is often stated that
we have lost the phase and hence the signal cannot be recovered. This argument
is not correct, because the phase we have lost is the phase of the short-time Fourier
transform, not the signal. In fact, the phase and amplitude of the signal appear in
both the phase and amplitude of the short-time Fourier transform. Therefore, hav-
ing the amplitude of the short-time Fourier transform may be sufficient to recover
the signal. We now see under what conditions this is the case. From Eq. (7.11) the
ambiguity function of the signal is
A8(9,r) = MsP(0, (7.90)
Ah(-B,
1 p) I2ejet+iT4, dt dw
-,9) (7.91)
Ah(
T I st
But the ambiguity function determines the signal. To see this, we Fourier invert the
ambiguity function of the signal, Eq. (7.12), to obtain
s(t) =
27rs (0)
1
f MsP(B, t) e_jet/2
Ah (-0, t)
d9 (7.94)
which shows that the signal can be recovered from the spectrogram. However,
there is an important caveat. The division by the ambiguity function of the window
must be possible and that will be the case if it has no finite regions where it is zero
in the 0, T plane. That depends on the window function.
Sec. 10 Expansion In Instantaneous Frequency 109
fA2fA2()+t)A t)dT
(7.95)
(w)t = ()d
Expanding go'(T + t) as a power series in r
For n = 0 the denominator and numerator are equal, and hence Mo(t) is always
equal to one. Writing out the first few terms for (w )t we have
Similar considerations lead to the following expansions for the mean square and
conditional standard deviation,
00 n (n-k+1) go (k+1)
(w2
)t
(w2
)t + E En=o k=0
(n - k)! k! Mn(t) (7.100)
00 n -1
21t = go(n-k+1) go(k+l) Mn - Mn-k!k
or (U)2)0
+ 1: -
-i 1
n=2 k=1
(n - k)! k! (7.102)
(M3-M2M1)Wit
ti (w2 )t + (M2-M12)go112 +
g" (7.103)
110 Chap. 7 The Short-Time Fourier Transform
Note that the derivative of the phase of the signal is absent from the expansion of
o ,t. Also, M2 - M1 is the local duration as defined by Eq. (3.34),
Tt2 =M2-Ml 2
(7.104)
s(t) _ (a/7r)'14
e-Qt2/2+9w(t) A2 (t)
= (a/x)1/2 a-ate (7.105)
(W )t = 4P -
a + aV
+() (-2+ a (pi
rr
(7.106)
+ a)
It
2 1
2(a+a) + 1
2(a+a)
112 _
(a+a)2
at it Ill 1
+2(a+a)
1
2 + a+a
a2t2 rr rrrr
(7.107)
and therefore
N 1 + T2 lp1,2(t) (7.110)
c
I
t 4T2
The minimum is achieved at
TtzlII"
t' N 2 cp" (t)
At that value the local standard deviation is
rr" I t ^' 'P"(t) (7.112)
Sec. 11 Optimal Window 111
We emphasize that 7 is not the width of the window. It is the width of the
modified signal at time t, which involves both the window and the signal. However,
if the signal is purely frequency modulated then T2 is the width of the window, in
which case we have
Optimal width of window N 1 (7.113)
21 V11(t)
[ for purely frequency modulated signals ]
which gives
amm = 101 -a (7.115)
Note that am,,, does not have to be positive. If it is negative the window will not be
normalizable but nonetheless the STFT will be normalizable as long as ao,m + a is pos-
itive. Hence divergent windows may be used and indeed are desirable and necessary
to obtain a high resolution. The standard deviation of the window is o,w = 1/(2a)
and we have that the optimal window is
O'w (7.116)
2am,,, 2(1,61 -a)
If we have a constant amplitude signal ( a = 0) then, ao,i,, = 1,3 1 in accordance with
Eq. (7.113).
e -t2/(2.,2
2
(7.118)
Ah(t) 2ittrw
We obtain
1 N
2C16"2
+ 3c26"
amm =
2Qw a+ ,p"4 + 6c1cp" + 12c2 (7.119)
where
"i + 8 "2 - at 40 will
Cl = 4 c (7.120)
C2 = 1 a2t2 [ Wit "i + 112 1
(7.121)
2 J
For a chirp where WP = pt, c1 and c2 are both zero and we have that am. = I R - a,
which agrees with Eq. (7.115).
112 Chap. 7 The Short-Time Fourier Transform
2ryt-3a-47t+(4-b )t t-_ 00
dun N
(272)1/3 -a+ 2cx2 - 'S
_r2)1/3
+3 (. )1/3
J
t1 t-0
(7.122)
Chapter 8
8.1 INTRODUCTION
The Wigner distribution' is the prototype of distributions that are qualitatively dif-
ferent from the spectrogram. The discovery of its strengths and shortcomings has
been a major thrust in the development of the field. Although it has often been
studied in contrast to the spectrogram, we will see in Chapter 9 that both are mem-
bers of a general class of representations.
Wigner was aware that there were other joint densities but chose what has now
become the Wigner distribution "because it seems to be the simplest". Nonetheless,
Kirkwood,[30'] a year later, and Terletsky,15411 a few years after, came up with a sim-
pler candidate that is commonly called the Rihaczek or Margenau-HiU distribution.
The Wigner distribution was introduced into signal analysis by Ville,[16'1 some
15 years after Wigner's paper. Vile gave a plausibility argument for it and "de-
rived" it by a method based on characteristic functions. Remarkably, the same type
of derivation was used by Moyal[390l at about the same time. These derivations
appeared to be from first principles; however, they contained an ambiguity and
' Wigner's original motivation for introducing it was to be able to calculate the quantum correction to
the second virial coefficient of a gas, which indicates how it deviates from the ideal gas law. Classically,
to calculate the second virial coefficient one needs a joint distribution of position and momentum. So
Wigner devised a joint distribution that gave, as marginals, the quantum mechanical distributions of
position and momentum. The quantum mechanics came in the distribution, but the distribution was
used in the classical manner. It was a hybrid method. We discuss the analogy of quantum mechanics to
signal analysis in Section 13.13. Also, Wigner was motivated in part by the work of Kirkwood and Mar-
genau who had previously calculated this quantity but Wigner improved on it. Kirkwood subsequently
developed what is now the standard theory for nonequilibrium statistical mechanics, the BBGKY Hier-
archy (the theory was developed independently by Bogoliubov, Born and Green, Kirkwood, and Yvon).
Kirkwood attempted to extend the classical theory to the quantum case and devised the distribution
commonly called the Rihaczek or Margenau-Hill distribution to do that. Many years later Margenau
and Hill derived the Margenau-Hill distribution. The importance of the Margenau-Hill work is not the
distribution but the derivation. They were also the first to consider joint distributions involving spin.
113
114 Chap. 8 The Wigner Distribution
the generalization of the method, which we shall call the characteristic operator
method, can be used to derive all distributions. This is discussed in Chapters 10
and 17. In a comprehensive study, Mark[J pointed out the cross term issue of
the Wigner distribution and also showed the relation of the Wigner distribution
to the spectrogram. In 1980, in an important set of papers, Claasen and Meck-
lenbraukeri11e,119,1201 developed a comprehensive approach and originated many
new ideas and procedures uniquely suited to the time-frequency situation.
Calculating properties of the Wigner distribution is fairly straightforward. How-
ever, they can be easily determined by the methods we will develop to study the
general class of distributions in Chapter 9. Hence in this chapter we do not delve
into detailed proofs for every property, but emphasize the interpretation of the re-
sults.
1
S' (w + 10) S(w - a-j'8 dB (8.2)
21r 29)
The equivalence of the two expressions is easily checked by writing the signal in
terms of the spectrum and substituting into Eq. (8.1). The Wigner distribution is
said to be bilinear in the signal because the signal enters twice in its calculation.
Notice that to obtain the Wigner distribution at a particular time we add up
pieces made up of the product of the signal at a past time multiplied by the signal
at a future time, the time into the past being equal to the time into the future. There-
fore, to determine the properties of the Wigner distribution at a time t we mentally
fold the left part of the signal over to the right to see if there is any overlap. If there
is, then those properties will be present now, at time t. If this simple point is kept
in mind, many issues and results regarding the Wigner distribution become clear.
For example, suppose we have a finite duration signal with noise appearing only
for a small part of the time. (We use noise only for the purpose of illustration and
our remarks hold for any other property of the signal.) Now let us pick a time and
ask whether noise will appear at that time. Fold the signal at that time and if in
the folding the noise is overlapped then noise will appear at the present time even
though there is no noise in the signal at this time. This is illustrated in Fig. 8.1.
Everything we have said for the time domain holds for the frequency domain
because the Wigner distribution is basically identical in form in both domains. An-
other important point is that the Wigner distribution weighs the far away times
equally to the near times. Hence the Wigner distribution is highly nonlocal.
Range of the Wigner Distribution. The above argument shows that for an infinite
duration signal, the Wigner distribution will be nonzero for all time, since no matter
Sec. 2 The Wigner Distribution 115
ta tb
w
Fig. 8.1 An easy way to ascertain the behavior of the Wigner distribution is to men-
tally fold over the signal about the time being considered and determine whether
the overlap includes properties of interest. We have illustrated a finite duration sig-
nal where noise appears for a small time interval. At time to no noise will appear
in the Wigner distribution because if we fold over the signal about to there is no
noise in the overlap. Now consider time tb. Folding over does result in overlap
with the noise and thus noise will apear at time tb even though there is no noise at
that time. Although we have used noise for illustration, the same arguments apply
to any other property.
tl t2 to t3
Fig. 8.2 The Wigner distribution is not generally zero when the signal is zero and
similarly it is not zero for values of frequency at places where the spectrum is zero.
To see this, consider the signal illustrated, which is zero for the interval (t2, t3). Now
focus at t = ta. Mentally fold over the signal and note that we have an overlap.
Hence the Wigner distribution will not be zero at to even though the signal is zero
then.
what time we choose, the folding of the right with the left parts will result in a
nonzero answer. Now let us consider a signal that has a beginning in time and
calculate the Wigner distribution for a time before the signal starts. Mentally folding
over the right with the left we will get zero, since there is nothing to the left to fold.
Hence the Wigner distribution will be zero for times before the signal starts. Also,
for a signal that stops, the Wigner distribution will be zero after that time. For a
finite duration signal the Wigner distribution will be zero before the start of the
signal and after the end. The same arguments in the spectral domain show that for
a bandlimited signal the Wigner distribution will be zero for frequencies outside
the band. Therefore, the Wigner distribution satisfies the finite support properties
in time and frequency,
W(t, w) = 0 fort outside (t1, t2) if s(t) is zero outside (t1, t2) (8.3)
W(t, w) = 0 for w outside (w1i w2) if S(w) is zero outside (w1i w2) (8.4)
116 Chap. 8 The Wigner Distribution
Now consider a signal that is turned off for a finite time and then turned on and
let us focus on a time for which the signal is zero, as illustrated in Fig. 8.2. Will
the Wigner distribution be zero? No, because if we fold over the right side with
the left side we do not obtain zero. Similar consideration applies to the spectrum.
Therefore, generally, the Wigner distribution is not necessarily zero at times when
the signal is zero and it is not necessarily zero for frequencies that do not exist in the
spectrum. Manifestations of this phenomenon have sometimes been called inter-
ference or cross terms and the cause for this behavior has very often been attributed
to the fact that the Wigner distribution is bilinear in the signal. It is not bilinearity as
such that is doing it since there are other joint distributions that are bilinear, satisfy
the marginals, but are always zero when the signal is zero. Particular illustrations of
this effect are presented in the examples and also when we discuss multicomponent
signals.
M(O, T) = if ejot+.7T'd
W (t, w) dt dw (8.5)
This function and variants of it have played a major role in signal analysis. This
particular form is called the symmetric ambiguity function. It was first derived by
Vile and Moyal and its relation to matched filters was developed by Woodward.
We have previously discussed it in the calculation of the characteristic function of
the spectrogram, Eq. (7.12). In terms of the spectrum the characteristic function is
M(9, T) = f S (w + 2e)
S(w -
Z9)
e3" dw (8.10)
Reality. The Wigner distribution is always real, even if the signal is complex. This
can be verified by considering the complex conjugate of W (t, w),
2_'r
_ 1
1
. -OO
s(t+2T)s*(t _ 2T)a-jT"dT (8.12)
27r
f O
s(t+ -IT) s*(t - 2T) a-'T" dr (8.13)
W(t, w) (8.14)
The fact that the Wigner distribution is real for any signal can also be seen from the
characteristic function. Recall that M* (-0, -T) = M(0, T) is the condition for a
distribution to be real. But the characteristic function of the Wigner distribution is
the ambiguity function, A(0, T), Eq. (8.8), which does satisfy this property.
Symmetry. Substituting -w for w into the Wigner distribution we see that we ob-
tain the identical form back if the signal is real. But real signals have symmetrical
spectra. Therefore, for symmetric spectra the Wigner distribution is symmetrical in
the frequency domain. Similarly for real spectra the time waveform is symmetrical
and the Wigner distribution is symmetric in time. Therefore,
W(t, w) = W(t, -w) for real signals - symmetrical spectra, S(w) = S(-w) (8.15)
W (t, w) = W(-t, w) for real spectra - symmetrical signals, s(t) = s(-t) (8.16)
Both of these equations can be readily verified by examining M(8, 0) and M(0, T).
By inspection of Eq. (8.8) and Eq. (8.10) we have
M(0, 0) = f (8.19)
118 Chap. 8 The Wigner Distribution
But these are the characteristic functions of the marginals and hence the marginals
are satisfied. To do it directly,
I s(t) 12
(8.22)
and similarly for the marginal in frequency. Since the marginals are satisfied, the
total energy condition is also automatically satisfied,
Tune and Frequency Shifts. If we time shift the signal by to and/or shift the spec-
trum by wo, then the Wigner distribution is shifted accordingly,
if s(t) - eJ"ots(t - to) then W (t, w) - W (t - to, w - wo) (8.24)
To see this we replace the signal by ejwot s(t - to) in the Wigner distribution and
call Wah the shifted distribution,
e-io(t-r/2)
Wsh(t,w = 1
s*(t - to - 2r)
21r J
x e wo(t+r/2) s(t - to + 2r) a-Jr" dr (8.25)
1
s (t - to - 2r)
1
s(t - to + 2r) a -7r("-"o) d7-
1
(8.26)
27r
Mean Time, Mean Frequency, Duration, Bandwidth, and the Uncertainty Princi-
ple. All these quantities are automatically satisfied since all these quantities depend
on the marginals only, which are correctly given by the Wigner distribution.
Correlation and Covariance. The first mixed moment of the Wigner distribution
is
(tw) = fftwW(tw)dtdi. = Jtw'(t) I s(t)12dt (8.31)
Therefore the Wigner distribution gives the covariance of a signal as defined by Eq.
(1.124).
where 0 and 0 are the phase and spectral phase of the signal. But W'(t) and -0'(W)
are the instantaneous frequency and group delay. These are important results be-
cause they are always true for any signal. Recall that for the spectrogram they were
never correctly given, although we could approximate one or the other. Further-
more, if we tried to make one of the two relations approximately true by narrowing
the window, the other relation would become very poor.
Local Spread. The result just obtained shows that instantaneous frequency is an
average, the conditional average for a particular time. We now consider the spread
about that average, the conditional standard deviation. First consider the second
conditional moment in frequency,
( w2
)t = s(t) fw2 W (t' w) dw (8.34)
12
1 2 - A"(t) + 2(t)
(8.35)
2 A(t)
CAI (t) / A(t)
120 Chap. 8 The Wigner Distribution
This expression for o may go negative and hence cannot be properly interpreted.
In fact, for such a case wIt is imaginary. Therefore, while the Wigner distribution
gives an excellent result for the average conditional frequency, it gives a very poor
one for the spread of those frequencies. Are there distributions that give a plausible
answer? The answer is yes and we consider this question in Section 13.2.
8.6 EXAMPLES
Before proceeding with the general properties of the Wigner distribution we con-
sider some examples.
For a sinusoid the Wigner distribution is totally concentrated along the frequency of
the sinusoid, and for the impulse it is totally concentrated at the time of occurrence.
We see that the Wigner distribution is still concentrated around the single frequency
of the sinusoid but now there is spreading. In addition, for any particular value of
frequency it falls off as the squared amplitude of the signal. To calculate the conditional
spread, Eq. (8.37), we have
Note that this expression for the conditional spread as given by the Wigner distribu-
tion is always the case for a Gaussian envelope since the conditional spread does not
depend on the phase of the signal. However, other distributions will give different
answers. See Section 13.2.
Sec. 6 Examples 121
The instantaneous frequency is w, = pt + wo and therefore, for this case, the Wigner
distribution is totally concentrated along the instantaneous frequency.
e-at'/2+j9t2/2+jWOt
s(t) = (a/7r) 1/4 W(t,w) = 1 e-ate-(W-9t-WO)'/a (8.43)
7r
This is the most general case for which the Wigner distribution is positive and this case
encompasses all previous examples. If a is small, then the distribution is concentrated
along the w = wo + ,Ot, which is the instantaneous frequency. In Fig. 8.3 we plot the
Wigner distribution.
5 -5
Fig. 8.3 The Wigner distribution for the signal s(t) = (a/7r) 1/4 e-at2 /2+j8t2 /2+jwot
This is the most general signal for which the Wigner distribution is positive through-
out the time-frequency plane. As the envelope becomes more and more flat the
Wigner distribution gets more and more concentrated along the instantaneous fre-
quency, as shown in (b). In (a) and (b) a = 1 and a = 0.001 respectively. In both
figures wo=8andf=1.
s(t) = (4a3/7r)1/4te-at'/2+j9t2/2+jWot
(8.44)
5 -5
Fig. 8.4 The Wigner distribution for the signal s(t) = t Ev-
erything below the plane is negative.
we have
2r) + p(t + 17-) = yr3/12 + ('yt2 + pt + wo)r (8.48)
Unlike the quadratic case, this distribution is not a delta function of w-cp' (t). However,
it is concentrated along w = cp' (t).
has an instantaneous frequency given by w; = /3t + wo + mwm cos w,t. The Wigner
distribution can be obtained analytically. Substituting the signal into the definition of
the Wiper distribution we obtain
W (t, w)
CIO f e-are/a+2am cos(wmt) sin(wmr/2)-7r(w-wo-pt) dr
tar (Ct ) 1/2
(8.57)
Now 00
e2jmcos(wmt) sm(wmr/2) ejnw-r/2 (8.58)
_ E Jn(2m cos(wmt))
n=-m
where J. is the Bessel function of order n. Substituting this into equation Eq. (8.57)
and integrating we have
00
This can be interpreted as the Wigner distribution of the sum of Gaussians each with
instantaneous frequency w, = wo + n'.&m /2 +,3t.
Complex Signals and Real Signals. The reader may have noticed that all the ex-
amples presented thus far have been for complex signals. The reason is that there
is an effective way to think of the Wigner distribution for the sum of two signals. A
124 Chap. 8 The Wigner Distribution
real signal can always be thought of as the sum of a complex signal plus its complex
conjugate. We therefore leave the consideration of real signals until we develop
some simple results regarding the Wigner distribution for the sum of two signals.
where
W12(t,w) = 2- fs(t_ 2T)82(t+ 2T)e--77"dr
(8.62)
W12 (t, w) =
2-
fSi(+ z 9) S2 (w - z 0) a-'te d9 (8.63)
The cross Wigner distribution is complex. However, W12 = Wzi, and therefore
W12 (t, w) + W21 (t, w) is real. Hence
We see that the Wigner distribution of the sum of two signals is not the sum of
the Wigner distribution of each signal but has the additional term 2 Re { W12 (t, w) }.
This term is often called the interference term or the cross term and it is often said to
give rise to artifacts. However, one has to be cautious with the images these words
evoke, because any signal can be broken up into an arbitrary number of parts and
the so-called cross terms are therefore not generally unique and do not characterize
anything but our own division of a signal into parts. Sometimes there is a natural
decomposition where the self terms and cross terms take on special meaning. This
question is addressed in the next section and in Section 13.3.
This is an illustration of the cross term. It is sharp at w = 1 (wi + w2) because that is
the only value for which there is an overlap, since the signal is sharp at both wi and
W2.
W (t, w) = 6(w - wo) + 6(w + wo) + 2 6(w) cos 2wot [ s(t) = cos wot ] (8.67)
Straightforwardly we obtain
W (t' w) = 6(w - wo) + 6(t - to) + cos[2(w - wo)(t - to) - woto] (8.69)
2=
Vr7r
Comparing this equation to the spectrogram, Eq. (7.69), we see that the self terms
are infinitely sharp - something the spectrogram cannot achieve. However, the cross
terms in the Wigner distribution are not hidden by the self terms as is the case in the
spectrogram, as per the discussion after Eq. (7.69). In Fig. 8.5 we plot this distribution.
We also give the ambiguity function for this case,
where c.c. stands for the complex conjugate of the preceding terms.
0 0
126 Chap. 8 The Wigner Distribution
where the modulation extends over all time, the cross terms will extend over all times
and frequencies. The Wigner distribution can be calculated exactly,
A1A2 2(alaa)1/2 r 2t 1
+2 a cos (w + w2a1 - W 1 0 2 )
a1 + a2 L a1 + az
and taking the specific value, t = 7-/2, letting k = 1/s` (0), and then setting r = t,
we have
s(t) = k f W (t/2, w) eat`'' dw (8.74)
which shows that the signal can be recovered from the Wigner distribution up to a
constant. The constant can be obtained from the normalization condition up to an
arbitrary constant phase factor. A constant phase factor in the signal clearly drops
out when we calculate the Wigner distribution, since we are multiplying the signal
times its complex conjugate. Therefore it can never be recovered.
Representability. Not every function of time and frequency is a proper Wigner dis-
tribution because there may not exist a signal that will generate it. If a two dimen-
sional time-frequency function is generated from some signal, we shall say that it is
128 Chap. 8 The Wigner Distribution
For what values of a, 0,,y can this be a Wigner distribution? Assume that it is and
obtain the signal using the inversion formula, Eq. (8.74),
(8.77)
When does W (t, w) = G(t, w)? Only when c = /i or
ry2
1 =1+ (8.78)/
aQ
Therefore the three constants must be chosen according to Eq. (8.78) if G(t, w) is to be
a Wigner distribution.
If the above calculation is repeated we see that there is no way to choose the constants
appropriately. There is a much easier way to ascertain that. Note that this G is always
nonnegative. We know that cannot be unless the distribution is given by Eq. (8.42),
and this G is not of that form. Hence it is not a representable Wigner function.
Overlap of Two Signals - Moyal Formula. In many problems the overlap of two
signals enters. It can be expressed as the overlap of their respective Wigner distri-
butions,
2
This is easily verified and was first shown by Moyal.13901 Of course, if we have two
signals it would make no sense to calculate their overlap by first calculating their
respective Wigner distributions and then doing the integrals indicated. However,
this expression is of theoretical interest and also has been applied to problems of
detection by Flandrin.11911
Range of the Cross Terms. We discussed the range of the Wigner distribution in
Section 8.2. Similar considerations establish the range of the cross Wigner distribu-
tion. In particular, for two functions, Si (t) and 32 (t) , which are respectively zero
outside the intervals (t1, t2) and (t3, t4), the cross Wigner distribution satisfies the
followingl2531
if s1(t) is zero outside (t1, t2) and s2(t) is zero outside (t3, t4)
In the frequency domain, for two bandlimited functions, Sl (w) and S2 (u)), which
are zero outside the intervals (wl, w2) and (w3i w4), respectively, we have
if Sl (w) is zero outside (w1, w2) and S2 (u;) is zero outside (w3, w4)
Product of Two Signals and Product of TWo Spectra. The Wigner distribution of
the product of two signals s(t) = si (t) s2(t) can be written in terms of the Wigner
distribution of each signal. The way to do it is to substitute S1 (t) 82 (t) into the defi-
nition of the Wigner distribution and then use the inversion formula for each signal.
This results in
W (t, w) = Wl (t, w') W2 (t, w - w') dw' [for s(t) = 81(032W1 (8.83)
J
Similarly, if we have a spectrum that is the product of two spectra (in which case
the signal is the convolution of the two signals), then
Analytic Signal. Since the analytic signal eliminates the negative frequencies, the
Wigner distribution of an analytic signal will be zero for the negative frequencies.
Also, use of the analytic signal will eliminate the cross terms between the negative
and positive parts of the spectrum.
distribution of the signal with the Wigner distribution of a window, we get the spec-
trogram
2
1
e-3Wrs(r) h(r - t) dr- (8.86)
Noise. The Wigner distribution is very noisy and even places noise at times where
there was no noise in the signal. To understand this, consider an infinite duration
signal where there was noise for only five minutes. Take any particular time and,
using our usual argument, fold the future with the past. The five minutes of noise,
no matter when it occurred, will contribute and therefore noise will appear even
though there was no noise at the time being considered. This will hold true for any
time. For a finite duration signal it will appear at those times when the folding over
procedure includes the noise.
Fig. 8.7 illustrates these effects. An analysis of additive white noise and the
Wigner distribution has been done by Nuttall. I43]
t1
frequencies below and above that frequency. There are two reasons for wanting
to modify this basic property of the Wigner distribution. First, in practice we may
not be able to integrate from minus to plus infinity and so one should study the
effects of limiting the range. Second, in calculating the distribution for a time t, we
may want to emphasize the properties near the time of interest compared to the
far away times. To achieve this, note that for a given time the Wigner distribution
is the Fourier transform with respect to r of the quantity s* (t - 1 r) s(t + 2 r). The
variable r is called the lag variable. Therefore if we want to emphasize the signal
around time t, we multiply this product by a function that is peaked around r = 0,
h(r) say, to define the pseudo Wigner distribution
In the non windowed version, Eq. (8.36), the Wigner distribution was totally concen-
trated at w = wo. It was W(t,w) = 6(w - wo). That is no longer the case and the
broadness depends on the window size.
Example 8.16: Windowed Wigner for the Sum of TWo Sine Waves.
For the sum of two sine waves
Alej"lt + A2ejw2t
s(t) = (8.90)
with the same h(t) as above we have
2AiA2
+ cos[(w2 - wl)t] e-('-("1+-2)/2)2/(2a) (8.91)
27ra
If we choose a small a the cross terms can be made small. However, note that the self
terms get spread out.
132 Chap. 8 The Wigner Distribution
Substituting this L and the definition of the Wigner distribution into Eq. (8.92) and
integrating over w' results in
WSM(t,w)
27r ff e-(t'-t)/a-)9T2 2Tr) s(t' + 2rr)dt'dr (8.94)
(8.95)
Expanding e(0-1/0)xy/2 in a Taylor series we have
2
WSM (t, w) =
Y V' (0 - 1/a)n f(x)xn dx (8.%)
27r
n=0
_ 2n
n! f
If each term is positive then the sum will be positive. To assure that we must take
- 1/a > 0 or equivalently
aQ> 1 (8.97)
Notice that if af3 = 1, then only the first term survives. But with this condition L
is a Wigner distribution, as we showed in Eq. (8.78). In that case, we know from
Eq. (8.85) that the resulting distribution is the spectrogram. In the general case, Eq.
(8.96) is a sum of spectrograms.
Nuttall1403, 404] has generalized this by considering the smoothing function
e-t2/-W2/A-2ywt
L(t,w) = (8.98)
hand, the spectrogram often cannot resolve the components effectively. We men-
tion here that other distributions we will consider keep the desirable properties of
the Wigner distribution and have considerably better behavior with regard to the
undesirable properties. In Figs. 8.8, 8.9, and 8.10 we give a few examples contrast-
ing the Wigner distribution and spectrogram.
Contrast Regarding Global Averages and the Uncertainty Principle. Since the
Wigner distribution satisfies the marginals, it always gives the correct answers for
averages of functions of frequency or time and always satisfies the uncertainty prin-
ciple of the signal. On the other hand, the spectrogram never gives the correct an-
swers for these averages and never satisfies the uncertainty principle of the signal.
T
W
Fig. 8.8 Comparison of the Wigner distribution (bottom) with the spectrogram (top)
for two multicomponent signals.
Comparison of the Wigner Distribution with the Spectrogram 135
Sec. 11
(a) (U)
(Hz)
450
300
150
N (b)
Fig. 8.10 Aneurysm signal. About 5% of the population has a small spherical en-
largement of the wall of the cerebral artery (aneurysm). Some aneurysms emit a
sound. In the above figure are the Wigner distribution (a) and spectrogram (b) of
such a sound. (Courtesy of M. Sun and R. J. Sclabassi.)
Chapter 9
General Approach
and the Kernel Method
9.1 INTRODUCTION
The Wigner distribution, as considered in signal analysis, was the first example of
a joint time-frequency distribution that was qualitatively different form the spec-
trogram. The idea of the spectrogram crystallized in the 1940s. Independent of
that development, there was considerable activity in the 1940s, 1950s, and 1960s,
devising distributions which were similar in spirit to the Wigner distribution in the
sense that they satisfied the marginals, the instantaneous frequency condition, and
other desirable properties. Among the distributions proposed then (in signal analy-
sis and quantum mechanics) were the Rihaczek, Page, and Margenau-Hill. In 1966
a method was devised that could generate in a simple manner an infinite number of
new ones. 1' ) The approach characterizes time-frequency distributions by an aux-
iliary function, the kernel function. The properties of a distribution are reflected by
simple constraints on the kernel, and by examining the kernel one readily can as-
certain the properties of the distribution. This allows one to pick and choose those
kernels that produce distributions with prescribed, desirable properties. This gen-
eral class can be derived by the method of characteristic functions. In this chapter
we explain the method and the general ideas associated with it and give the deriva-
tion in the next chapter.
e-yet-jTw+jeu
C(t, w) = 471 JJ s* (u - 2 r) s(u 2 r) 0(9, r) du dr dB (9.1)
J
where q5(0, r) is a two dimensional function called the kernel, a term coined by
136
Sec. 2 General Class 137
Claasen and Mecklenbraukerl1201 and whom, with Janssenl2761, made many impor-
tant contributions to the general understanding of the general class, particularly in
the signal analysis context.
The kernel determines the distribution and its properties. In Table 9.1 we list
some distributions and their corresponding kernels. For the Wigner distribution
the kernel is one; however, no particular significance should be attached to that
since it is possible to write the general form so that the kernel of any distribution is
one, in which case the kernel of the Wigner distribution would be something else.
as may be verified by expressing the signal in terms of the spectrum and substituting
in Eq. (9.1).
Alternate Forms. There are a number of alternative ways of writing the general class
of time-frequency distributions that are convenient and add considerable physical
insight.
where
M(O"r) = 00, T) s* (u - 2 T) s(u + 2'r) a ou du (9.4)
J
= 0(9, r) A(0, -r) (9.5)
and where A(O, T) is the symmetrical ambiguity function. The characteristic func-
tion may be appropriately called the generalized ambiguity function.
Rt(T)
_ JJ 8*(u - 2T) s(u + 2T) 0(9, r) e'e("-t0 d9 du (9.7)
Wigner[584] 1
1 e-""Ws'(t - 1r)s(t+ 1r)dr
27r 2 2
Margenau-Hill[358] cos BT
Z
Re ---s(t)S(w)e"
27<
e7elrl 8
Page [4191
at
Choi-Williams[117] e-B2r2/
47x3/2 f T2/Q
a' (u- 2T)s(u.+2T)dudr
Zhao-Atlas-Marks [626]
9(T)ITI aB
8r _f1
47ra
g(r)e-)r"
1 Derived in reference [125] using the ordering rule of Born and Jordan. The derivation is repeated in
Section 10.6.
2 These distributions and the spectrogram are manifestly positive, but only the 'Positive' ones satisfy the
margin's.
Sec. 2 General Class 139
Fourier Transform of the Kernel. Notice that in the general form 9 does not appear
in the signal. Hence if we define
r(t, T) = 27r
1 0(9, r) a-.it9 d9 (9.8)
Rt(T) fr(t_u,T)s*(u_T)s(u+r)du
2 (9.10)
then we have
C(t, W) = Jf K(t, w; x', x) s` (x) s(x) dx dx' (9.12)
with
K(t, w; x', x) = 2 r(t - (x + x'), - x) e-jw(x'-x)
(9.13)
2
As we discuss in the subsequent sections, properties of the distribution are reflected
in the kernel. These constraints may be imposed on 0 or K but are generally sim-
pler when expressed in terms of 0. This is a bilinear form only when K is signal
independent.
the distribution is
'Types of Kernels. An important subclass are those distributions for which the ker-
nel is a function of the product Or,
For notational clarity, we will drop the subscript PR since we can tell whether we
are talking about the general case or product case by the number of variables at-
tributed to 0(0,,r). Kernels that are a product of a function of each variable
0(0,'r) = 1(9) t2(r) separable kernels (9.18)
Bilinearity. If the kernel does not explicitly depend on the signal then the signal
enters into the general form twice, hence one says that the distribution is bilinear, a
phrase used by WigneL All the distributions listed in Table 9.1 are bilinear since the
kernel does not depend on the signal explicitly. Distributions obtained by taking
the kernel to be signal dependent are discussed in Sections 13.5 and Chapter 14.
In this chapter we will assume that the kernel is bilinear and time and frequency
independent. However, the methods developed for the bilinear case often apply
directly to the nonbilinear case. We point that out when appropriate. Notice that
Sec. 4 Basic Properties Related to the Kernel 141
the marginals are bilinear in the signal. However, that in no way reflects on the
functional dependence of the joint distribution. Having bilinear marginals does
not imply that the joint distribution must be bilinear
Determination of the Kernel. If a distribution is given we can usually pick out the
kernel by putting the distribution in one of the standard forms. If that is not readily
doable, we can calculate it by finding the characteristic function of the distribution
and using Eq. (9.5). Explicitly,
C(t, w) dw = 2
J - fff b(r) s' (u - 2 T) s(u + 2T) 0(9, T) eJ8(u-t) d9 du d7- (9.21)
12 &9(u-t) dO du (9.22)
= 27r
ff 0(e, 0) 1 s(u)
If we want this to equal I s(t) 12 then the 9 integration must be made to give
1
(B, 0) e' B(u-t) dB = b(t - u) (9.23)
27r
142 Chap. 9 General Approach and the Kernel Method
This is the condition for the time marginal to be satisfied. Similarly, for the fre-
quency marginal to be satisfied
we must take
0(0,7-) = 1 [ for frequency marginal 1 (9.26)
Total Energy. If the marginals are correctly given, then of course the total energy
will be the energy of the signal. However, we can retain total energy conservation
(normalization to one) without insisting on the marginals. Integrating the general
form with respect to time and frequency shows that for the total energy to be pre-
served,
Jf C(t, w) dw dt = 1 [ for total energy ] (9.27)
Reality. For a distribution to be real the characteristic function must satisfy M(8, T) =
M*(-0, --r). Look at Eq. (9.5) for the characteristic function and note that the
ambiguity function satisfies A(9,T) = A*(-9, -T). Therefore the only way that
M(9, T) = M* (-0, -T) is for the kernel to also satisfy the identical condition,
Time and Frequency Shifts. A signal that is shifted by a time to and frequency
wo is sah = ei(0ts(t - to), where to is the amount of time translation, and wo is
the translation in the frequency domain. Substituting this into the general form we
have, using Cah for the translated distribution,
Sec. 4 Basic Properties Related to the Kernel 143
C'ah(t,w) ejwo(u+r/2-to)
= T7-r2
1 / //
4'
fJf (o, r)s'(u - 2T) s(u + 2T) e-jB(t-to)-jr("-"o) +ie+' do d-r du (9.33)
47r2
In going from step (9.31) to step (9.32) we assumed that the kernel is not a function
of time and frequency. Therefore
Scaling Invariance. In Section 6.5 we showed that if a signal is linearly scaled then
the spectrum is inversely scaled and we argued that a plausible requirement on the
joint distribution is that it scale the same way in time and frequency. In particular,
if C., (t, w) is the distribution of the scaled signal, s,,,, then the requirement is
C`(t' w) N sf (a(u - 2r)) s(a(u + 2-r)) 0(0, rr) e-set-3rw +,9u do d7- du
4a2
(9.38)
fransforming by u -+ u/a, T -+ T/a, and o -- ao, we have
C,C(t, w)
47r2
fJf sf (u - 7) a(u + 2r) O(ao, 7-/a) +jOu do dT du
(9.39)
We see that
The only way that a function of two variables can satisfy 0(a9, z/a) = 0(0, r) is if it
is a function of the product of the two variables. Therefore, for scale invariance the
kernel must be a product kernel,
Note that we have implicitly assumed that the kernel is signal independent.
Weak Finite Support Recall from Section 6.6 that weak finite time support means
that for a finite duration signal we want the distribution to be zero before the signal
starts and after the signal ends. The condition for that to be the case is
O(0,r)a-jetdO
=0 fori7I < 21tj (9.42)
J
[ for weak finite time support ]
Similarly for a bandlimited signal we want the distribution to be zero outside the
band. Then condition for that to be so is
Strong Finite Support Strong finite support means that the distribution is zero
whenever the signal or spectrum is zero. The conditions for this to be the case were
derived by Loughlin, Pitton, and Atlas(] for the case of bilinear distributions. The
conditions are
7-)e-jT"
f 0(9, dT = 0 forIBI # 21wl (9.45)
Ocw(0, r) = e-92_2/0,
(9.46)
where v is a constant. It is clear that for either 0 or -r being zero the kernel is one and
hence the Choi-Williams distribution satisfies the marginals. Also, since it is time and
frequency independent it satisfies the shift properties. For the finite support property
we have
/'
J 0(0, r) a-jet dO =
T exp [-:;] (9.47)
Therefore it does not satisfy the weak finite support properties but does so approxi-
mately for v sufficiently large.
Sec. 4 Basic Properties Related to the Kernel 145
and this dearly satisfies the weak finite support property, Eq. (9.42).
sin(a9T)
I ti < or
f aBT
a-'Bt dt aT
0 otherwise
(9.51)
Comparing to the condition for weak finite support, Eq. (9.42), we see that the condi-
tion is satisfied only if a < i .
s(t) = 1
27rs*(0) f M(t7, t) e-,jOt/2
0(0, t)
d9 (9.55)
C(t', w) ejtw+je(t'-t/2)
N
1
dt' &.0 d9 (9.56)
27rs (0) (B, t)
Hence the signal can be recovered from the distribution if the division by the kernel
is uniquely defined for all values of 0,,r. If the kernel is zero in regions, we cannot
recover the signal uniquely. For the spectrogram this will depend on the window.
2as*
1(0) fff dt' dw dO (9.57)
If the marginals are satisfied, then averages of the form g(t) = g1(t) + 92 (w) will be
automatically satisfied,
Correlation Coefficient and Covariance. Let us now consider the first mixed mo-
ment,
82M(0,T)
(wt) = (9.63)
00197- 0,,r =0
Sec. 6 Local Averages 147
02
0(0, rr)
9,T - o f A2(t) dt + J t 0'(t) A2(t) dt (9.64)
If we want the covariance to be equal to the form given in Section 1.8 then we
must have that (tw) = (t cp'(t)), which can be achieved by taking the mixed partial
derivative of the kernel to be zero at the origin,
= 0 (9.66)
aOa-r 9,T = o
= e,T = 0
and we see that the Choi-Williams distribution gives the proper expression for the
covariance.
C(w It) =
C(t, )
C(t I w) =
C(t,
) (9.67)
where P(t) and P(w) are the marginal distributions. Thus the expectation value of
a function g at a given time or frequency is
Instantaneous Frequency and Its Spread. For the average frequency at a given time
we calculate
1
ff A 2(u) [(e , 0) (u) - (O
1aO(e, T)
T
a u-t)2 d0 du (9.71)
1
p(t) [ 0(0) A2(t) (t) + 2 0'(0) A(t) A'(t) ] [ product kernel ] (9.72)
148 Chap 9 General Approach and the Kernel Method
Suppose we satisfy the time marginal in which case P(t) = A2 (t). Then
190(0, T) &6(u-t) de du
(w )t = (t) - i Pit) 2 f f A2(u) (9.73)
r=0
If we want this local average to be equal to the derivative of the phase we must take
the integrand to be zero. Therefore
80(6, T)
(w )t = W'(t)
if =0 (9.74)
8z T=0
Second Conditional Moment of Frequency. For simplicity we give the result for
product kernels:
(0) ] A'2
(w2 )tP(t) = [ 0(0) + 4" (t) - 2 [ 0(0) - 40" (0) 1 A(t) A" (t)
z
+ 0(0) A2(t) 2(t) + 2'(0) [ 2 A(t) A'(t) p'(t) + A2(t) (P11 (t) ] (9.76)
If we impose the conditions to obtain the derivative of the phase for the first con-
ditional moment, that is, the marginal conditions and 0'(0) = 0, this becomes
(N(t) \2
(w2 )t = [ 1 + 40"(0) ] A(t) 1 - Z [1 - A11(t)
A(t) +'P (t)
'2
(9.77)
2
Standard Deviation. The local standard deviation is obtained from a'2'It = (w2 )t -
(w )2.. From Eqs. (9.77) and (9.74) we have
A' (t)
Q2 O = a [ 1 + 40"(0) ] A(t) J 2 - Z [ 1 - 40"(0) ]
(t)
Positive Spread. In general, the conditional second moment and standard devia-
tion become negative and cannot be interpreted. However, for the choice
(0) = 4 (9.79)
we obtain
(A,(t) \2
2
fl (9.81)
GO A(t)
Sec. 7 Transformation Between Distributions 149
which are manifestly positive. Are these plausible results? We discuss this issue in
detail in Section 13.3 where we develop the idea of instantaneous bandwidth and
connect it to the nature of multicomponent signals.
2
Q ,1t = 1 A'(t) 2 A"(t) (9.82)
2 A(t)) A(t)
Group Delay. If we want the conditional moment of time for a given frequency to
be the derivative of the spectral phase then the identical approach as above gives
M,(0,,r) _ 01 (0"r)
M2(9,T) (9.86)
02(e,T)
To obtain the relationship between the distributions take the double Fourier
transform of both sides and use Eq. (9.3) to obtain
Cl
(t ' w) 41 2
J f 01 (0, T)
M2 ( 0 , ,r) a -jet -.7 d8 d r ( 9 . 87)
C1
(t '
w) - fJJ 02(
= 47Ir2 J 0, r) C
,
2(ti ,w P
)e jB(t'-t)+jr(W'-W)dBd r dt'dw ' ( 9 . 88)
with
912 (t, w) = 1r J
01(9, r) bet+2TW
2 (0, r)
dB dr (9.90)
_ 1 Ah(-8, r) e3et+fr
9sp(t, w) d8 dr (9.92)
47r2 ff 0(0,7-)
X
0(0,,r) e-Set+.iTW+f eu
du dr d0 (9.94)
0(0"r) 0(-0, r)
If we take kernels for which 0(-9, r) 0(0,,r) = 1, we see that gsp(t, w) is just the
distribution of the window function, except that it is evaluated at -w. Therefore
and
Positive Bilinear Distributions. In Section 8.10 we saw that if the Wigner distribu-
tion is convolved with certain functions a positive distribution is obtained. We have
also just seen that when the general class is convolved with certain g's the spectro-
gram is obtained. The general question as to when one obtains a positive bilinear
distribution b convolving the general class has been comprehensively addressed
by Janssen[Jwho derived a number of important results relating to this question.
Chapter 10
Characteristic Function
Operator Method
10.1 INTRODUCTION
In the last chapter we showed that an infinite number of joint time-frequency dis-
tributions can be generated. We now describe a method for deriving them from first
principles. In this chapter we develop the ideas for time and frequency; in Chapter
17 we generalize to arbitrary variables and subsequently specialize to the physical
variable "scale".
and
C(t, w) fJ M(8, T) dO dT (10.2)
The aim is to find the distribution. If it were possible to find the characteristic func-
tion without first knowing the distribution, our problem would be solved because
we would then obtain the distribution by way of Eq. (10.2).
152
Sec. 2 Characteristic Function Method 153
We shall call M the characteristic function operator. What can we take for M? One
possible choice is to substitute the time and frequency operators in the expression
ejet+jT" for the time and frequency variables,
east+jrw ...., Mw = e2BT+jrW [ Weyli501 ] (10.6)
which is called the normal correspondence. Now, for ordinary variables ejst+jrw _
ejet ejr", but for operators it is not true that ejsT+jrW equals e'OT ejrw because the
operators do not commute. Since these two choices are not equal they will give
different answers for the characteristic function and therefore produce different
distributions. In addition to the above two choices the following are also equally
plausible,
ejOt+jr" _, elrw ejOT [ Antinormal 1 (10.8)
-4 [ e3OT e3rW + eJrW eyOT 1
2 [ Symmetrization ] (10.9)
sin 0-r/2 e3rW+joT J
[ Cohen[1J ] (10.10)
8r/2
As we will show, we can generalize these rules by
e?et+jr" -, 0(0,,r) &0T+,TW [ Coheni1211 1 (10.11)
154 Chap. 10 Characteristic Function Operator Method
There is one basic reason why all these correspondence rules are plausible and
equally correct in some sense. Recall from Eqs. (4.44)-(4.45) that the joint characteris-
tic function of two variables, x, y, is related to the individual characteristic functions,
M.(9), M. (,r), by
M.(9) = M(9,0) ; My(7-) = M(0,T) (10.12)
if we are to be assured of the right marginals. Moreover, this assures that averages
of functions of time or frequency only are correctly given. All the correspondences
given above satisfy these constraining equations.
To obtain a distribution we choose a characteristic function operator, calculate
the characteristic function by way of Eq. (10.5), and then find the distribution via
Eq. (10.2). These equations can be combined. Substituting Eq. (10.5) into Eq.. (10.2)
we have
where Tu and Wu signify the time and frequency operators operating on functions
of u. We now address the evaluation of expressions like Eq. (10.5) for the various
choices of the characteristic function operator.
= f S*(w)M(0,7-;T,W)S(w)dw (10.17)
If we are using the signal we will say that we are in the time domain, and if we are
using the spectrum we will say we are in the frequency domain. We recall from
Section 1.4 that the explicit forms of the time and frequency operators are
TW-WT=j (10.20)
The third is an important result of operator theory, which is well known but far
from trivial,
ejOT+jrW = e-jO7-/2 e;OT (10.22)
= ejer/2 e,OT ejrW (10.23)
We will not prove this here because we return to it in our considerations of a more
general case in Chapter 17.
In general,
We are now ready to see, by way of examples, how to derive distributions using the
characteristic function approach.
_ fs*(t)e9n/2e8TeJ1Ws(t)dt (10.26)
=
I
T7r ff ?(u - z 6(u - t) 8(u + 2 r) dr du (10.31)
Using the ordering given by Eq. (10.7) we have for the characteristic function
The distribution is
P(t ' w )
47r2
r fJ s' ( u) e jeu s(u + , r) CIO' -'T'" d9 dr du ( 10. 34)
1 't'S
s * ( t )e (w) ( 10. 35)
27r
If the correspondence of Eq. (10.8) is used, the complex conjugate of Eq. (10.35) is ob-
tained and if the symmetrization rule is used, Eq. (10.9), we obtain the real part.
The general characteristic function is the average of the characteristic function op-
erator
10.5 AVERAGES
If we have a function of time and frequency, g(t, w), its average can be calculated
using the time-frequency distribution or directly from the signal,
(g(t,w))
f
= ff g(t,w)C(t,w)dtdw (10.42)
where
r) e-jet-jT dt &,, (10.45)
7(0, = 47r2 f f g (t' w)
Or, equivalently,
To prove the equivalence of the two methods of calculating averages, Eqs. (10.42)
and (10.43), consider
e-jet-j` 0(e, r) eu
47r2 fJJJJ g (t, w)
ff(tw)C(tw)dtdw (10.51)
158 Chap. 10 Characteristic Function Operator Method
M(9, T) = V- (.7e)"(.7T)"`
(t.,wm)
n=0 m= n!! (10.52)
If we had the moments we could calculate the characteristic function and the distri-
bution. Moments are average values and hence should be calculable directly from
the signal by an operator that corresponds to the moments. We symbolically write
tnwm _ Cnm(T,W) (10.53)
In
2n E
B)T n-tW-Tt (10.59)
t=o J
m-tvwl [ Weyll 8] ]
= 2m E C )W (10.60)
t-o
'In manipulating these types of expressions the following relations are useful:
mm(m,n)
Wn'T"' = L -7)t ) (1)I (7)t T`-[ Wm-[
1=0
mm(m,n)
T"Wm = E 7l t! \ p) (e) Wm-[ ?-n-t
1=0
They were given, in the quantum mechanical context, by Born and Jordan. 1911.
Sec. 6 The Moment Method 159
1. T"_lWm i (10.61)
n+1 1=0
ej9t+jTW =
00
n=0mLL.=0
` (7e)r
CO n r )m
tnWn (10.64)
n=0 m=0
L "0
nn (T W) (10.65)
However, we know that the most general characteristic function operator can be
expressed in terms of the kernel, Eq. (10.36), and therefore
1 an+m
Cnm(T, W) = jm aenarm O(O, T) (10.69)
jn 9,r = 0
1 an+m drW
Cnm(T, W) = jnjm aenaTm (10.71)
8,T = 0
= TnWm (10.72)
We show that indeed Eq. (10.6) and Eq. (10.60) are equivalent. This was shown by
McCoy.! For the Weyl rule the kernel is 1 and
1
an+m
e39T+jrW
Cnn,(T, W) = BBn&r" (10.73)
9.T=0
= (M Wm-tmw (10.75)
2m
l=0
The step from Eq. (10.74) to Eq. (10.75) is not trivial but can be shown by using the
relations in the footnote of Section 10.6.
We consider the rule of Born-Jordan, Eq. (10.62), and calculate the characteristic func-
tion,
M 00
M(9, T)
EE (78)im!)m
n!! (C.". (T' W))
n=0 m=0
(10.76)
00 00 n
fs.(t)T11_twmrs(t)dt
Y- EE
n=0 m=0 t=0
(n + 1)n!m!
(10.77)
m n (,8)n
fsi(t)t_te2TWts(t)dt (10.78)
m+ 1n=0
1=0
where in the last step we have used the fact that e'rw is the translation operator. Now
n
[ [ _ (1 + r/t)"+' 1
E
t=o
(10.80)
and further
eiet(l+r/t) _ e Ot
to (je)n (1 +'r/t)n+1 _ 1
a (n + 1)n! r/t j0
(10.81)
M(B, r) s nO 22 ever/2
(10.83)
J s(t)Cets(t+r)dt
sin Br/2
8 (u - 2 r) eje's(u + 2 r) du (10.84)
Br/2
Therefore the kernel is
sin Br/2
(10.85)
8r/2
For a further discussion of this distribution see Section 12.4.
Chapter 11
11.1 INTRODUCTION
In our study of the Wigner distribution we saw that it sometimes gives values in
the time-frequency plane that do not correspond to our intuitive sense of how the
energy should be distributed. This is particularly the case when the signal is mul-
ticomponent. These spurious values are sometimes called interference terms be-
cause they happen whenever we have two or more components that "interfere."
It must be understood that the word interference as used in optics is a real effect,
but its usage here connotes something artificial and undesirable. A major develop-
ment in the field occurred when the conditions on the kernel that minimized these
spurious values were understood and implemented by Choi and Williams,1117) and
Zhao, Atlas, and Marks,11211 Loughlin, Pitton and L. Atlas,f4) Jeong and Williams,11]
and Cunningham and Williams.!"" In this chapter we use a simple but general ap-
proach to see what those conditions are and also describe how this has led to the
concept of kernel design.
162
Sec. 2 Reduced Interference Distributions 163
with
and
)12 = 2(wl + w2) (11.5)
The terms M22 and M21 are obtained from the above by interchanging the sub-
scripts 1 and 2. The distribution is
where
I2
C11 = I Al 0(0 T) e-.7T(w-wl) dr (11.7)
21r
with
K(w) = 2n f O(wl - W2,7-) e-jT(W-rV12) d7- (11.9)
To make the cross term small we must make K small. But K is a Fourier trans-
form with respect to r and to make a Fourier transform relatively flat the integrand
mustbe a peaked function. Therefore, as a function of r we must take 0(0, r) highly
peaked.
Everything we have said for the sum of two sine waves applies to the case of a
signal composed of two impulses,
because there is total symmetry in the general equation with regard to the signal
and spectrum. The effects discussed will be the same except that now we focus on
the 0 variable. Therefore if we don't want cross terms for the case of two impulses
then the kernel must be peaked as a function of 0. For the general case, when we
have two components that are not necessarily parallel to the time or frequency axis,
the kernel must be peaked in both 9 and r. We also know that the characteristic
function is given by M(8, r) = 0(9, 7-)A(9, r) and that the maximum is at the origin.
164 Chap. 11 Kernel Design for Reduced Interference
Hence the maximum of the kernel is also at the origin and we may take it to be one
at the origin since that is the requirement for the preservation of total energy. If we
also want the marginals to be satisfied, the kernel must be one along the 0 and -r
axes.
Therefore we seek kernels whose values away from either the 0 or r axis are
small relative to the value at the 0 and r axes. A way to describe this region is
to observe that the product O r is large when we are away_from either axis. We
conclude that for cross term minimization
0(0,rr) << 1 for Or > > 0 (11.11)
K =
1
I w2 _
wl
I
1
27r
j O(X)
a-jxv dx W-GJ12
(11.13)
Let us now see how some of the distributions we have considered thus far behave
with regard to the cross term property derived above.
Once the basic conditions for cross term minimization are understood it is easy to
find kernels that satisfy the basic properties and give finer control for certain signals.
One extension has been introduced by Diethorn,11751 and Papandreou and Boudreaux-
Bartels,l1
= e-(8/a)'(T/6)m
t(9, r)
This gives individual control in the 0, T directions.
In Table 11.1 we list the constraints for obtaining desirable properties on 0 and h(t).
The main advantage of formulating kernel design in h(t) is that we have been
able to collapse a number of diverse conditions into a few rather simple ones. In
addition, many methods that have been developed for filter design can be used to
produce such functions.
166 Chap. 11 Kernel Design for Reduced Interference
we ask whether, for any a between zero and one, 03 (0,,r) also satisfies that con-
straint? If it does, then the set of kernels satisfying this constraint form a convex
set. For example, consider the time marginal constraint, which is 0(0, 0) = 1. For
any two functions that satisfy this constraint, will 03 satisfy it? Consider
and we see that 03 (9, 0) satisfies the time marginal constraint. Hence such functions
form a convex set. The other constraints we have considered also form convex sets.
The method called "projection onto convex sets" automatically picks out the
functions, if they exist, that satisfy all the conditions. Furthermore, if such a function
does not exist then the method picks out the best function in the mean square sense.
This method was devised by Oh, Marks II, Atlas, and Pitton.l'i
ff O(9, r) 2 dr db a(11.21)
One chooses aand constrains the maximization of the functional f with this con-
straint. The bigger a, the more concentrated the kernel will be. In addition to the
above, we may also want to constrain the solution to give the correct marginals and
other quantities. An illustration is given in Fig. 11.1.
0
2.5
Time (msec.)
Some Distributions
12.1 INTRODUCTION
We have already discussed in detail the Wigner distribution and the spectrogram.
In this chapter we study additional distributions (some new, some old), and de-
scribe their properties and the physical and mathematical reasons for their intro-
duction.
168
Sec. 2 Choi-Williams Method 169
For large or
lim ll(w,W1,W2,a) = 6(w - 2(wl +w2)) (12.7)
o-00
and therefore the distribution will be infinitely peaked at w = 2 (wl + W2), which is
precisely the Wigner distribution. As long as a is kept finite, the cross terms will be
finite and not concentrated at one point. They spread out with a lower maximum
intensity. In Fig. 12.1 we illustrate the effect of different choices of a where the
delta functions are symbolically represented. Notice that the self terms are still delta
functions along the frequencies wl and w2. In general, this will not be true for other
signals and it is not generally true of the Wigner distribution either. It is true for the
Wigner distribution of a chirp only. For a chirp the Choi-Williams distribution is
e-p2T4/v e-3T(W-pt) dT
Ccw(t,w) = 27r f [s(t) = ept2/2) ] (12.8)
The concentration is along the instantaneous frequency and the width depends on
both ,3 and a. Figs. 12.1-12.6 illustrate the Choi-Williams distribution for a number
of analytic and real signals and in some cases we compare it to the spectrogram
and/or Wigner distribution.
WI
FREQUENCY -+
Fig. 12.1 The Wigner distribution (a) and Choi-WdUams distribution (b) and (c),
for the sum of two sine waves. Both distributions are infinitely peaked at the fre-
quencies of the two sine waves. In the Wigner distribution the cross terms also
go to infinity. In the Choi-Williams distribution their intensity is a function of the
parameter Q. In (b) we have used a large a, while in (c) a small a is used. The
Choi-Williams distribution becomes the Wigner distribution for a -* oo.
170 Chap. 12 Some Distributions
FREQUENCY
Fig. 12.2 A comparison of the Wigner (left) and Choi-Williams (right) distribution
for the sum of two chirps. (Courtesy of W Williams.)
(a) (b)
Spectrogram Binomial Transform
240
T
c 120
Q)
7
a)
LL
kh
-; i
(a)
3.8
(c)
0
Time (ms) 70
Time
Fig. 12.6 In some individuals the jaw makes a click when speaking. These sounds
are called TMJ sounds. Because of their very short duration the spectrogram (left)
does not properly reveal the time-frequency structure. On the right is the Choi-
Williams distribution. (Courtesy of S. E. Widmalm, W J. Williams and C. Zheng.)
2 ) WI
1 -(a+'2 ) W 2 1
dr (12.10)
e
and we see that K(w) is a function of w - w1 and w - w2, which are the location of
the self terms. Loughlin, Pitton, and Atlas[M11 have shown that for K to fall exactly
on the self terms, the kernel f (8, r) must produce a distribution whose cross terms
lie exactly midway between the self terms.
Depending on the choice of f (0,,r) these kernels may or may not satisfy other
desirable properties, such as the marginals. But nonetheless they can give a good
Sec. 3 Zhao-Atlas-Marks Distribution 173
indication of the time-frequency structure of a signal. Note that for two impulses
cross terms will be produced. However, for multicomponent signals that are mostly
parallel in the frequency direction the cross terms will not be prominent because
they will be hidden under the self terms.
A distribution with this property is the Zhao, Atlas, and Marks distribution,["]
OT
OZAM (0, 7) = 9('r) 1 T j s a9 (12.12)
which gives
r rt+aIr,
C(t, g(T) e-ir" s' (u - .17-) s(u + 2T) du d'r (12.13)
4a J Jt-airl
for the distribution. In the original work of Zhao, Atlas, and Marks, g was taken to
be equal to one and a to one half. This distribution has many interestin properties
and a comprehensive analysis of it has been given by Oh and Marks ] In Figs.
12.7 and 12.8 we show two examples.
(a)
(a)
(b)
/
the frequency band w to w + &. is
rW+n", 1 +o
i(t) = iW (t) dw 2x V", e'"'t dw (12.19)
J
1 Born and Jordan did not consider joint distributions. This distribution was derived in reference [125]
and called the Born-Jordan distribution because the derivation is based on the Born-Jordan rule. The
derivation is repeated in Section 10.6.
Sec. 6 Running Spectrum 175
=
f
t+t
V(t)i(t)dt =
1
27r
r
Jt JL
rE(t,w)
Power is the product of voltage and current V (t) i* (t) and is the energy per unit
time. Hence the energy in the time interval At and frequency interval Aw is
VV(t)e"tdwdt (12.20)
and therefore the energy density in time and frequency is
wherein the last step we have identified the signal with the voltage, s(t) = V (t),
in which case V,, = S(w). Equation (12.22) is the Margenau-Hill distribution de-
rived in Section 10.3, also called the Rihaczek distribution and complex energy spec-
trum. Although it does satisfy the marginals it does not have many other desirable
properties. In particular, it does not satisfy the instantaneous frequency condition,
although it does satisfy the strong finite support property.
f 00
P-(tt,w) dt' = I St (w) I2 (12.24)
This is the Page distribution. Its kernel is 0(0,,r) = eiO I T I /2 . By inspection we see
that it satisfies the marginals. The main characteristic of this distribution is that the
future does not affect the past and hence the longer a frequency exists, the larger
the intensity of that frequency as time increases. Once the particular frequency
stops, the distribution at that frequency remains constant as time increases. The
Page distribution satisfies the weak and strong finite support properties.
176 Chap. 12 Some Distributions
Fig. 12.9 A comparison of the Wigner distribution and Page distribution for a finite
duration sine wave. The Page distribution does not go to zero at the end of the
signal.
As time increases, the distribution becomes more and more peaked at wo. This is illus-
trated in Fig. 12.9.
Variations of the Page derivation were given by Turner 1M31 Levin,1332] and others.
Levin defined the future running transform by
oo
1 s(t') a-i"t' dt' (12.29)
St
(w) 27r Jt
and
s. (t) St ej"
P+(t, w) _ - at I Si (w) 12 = 2 Re
19
2-
(w) (12.31)
Note that if the two distributions are averaged we get the Rihaczek distribution.
Sec. 6 Running Spectrum 177
Similar to the running Fourier transform, we can define the running signal trans-
form for frequencies up to w by
8w (t)
= 7_n
2- f oo
S(w') -J-'t du)' (12.32)
Fig. 12.10 compares the Wigner, Page, and Margenau-Hill distributions for a
finite duration signal that has been turned on and off as illustrated.
FREQUENCY
Fig. 12.10 A comparison of the Wigner (a), Margenau-Hill (b), and Page (c) distri-
bution for a finite duration signal that is turned off and on as illustrated.
Chapter 13
Further Developments
13.1 INTRODUCTION
In this chapter we address a number of topics that are central to the description of
signals in time and frequency.
From this point of view, instantaneous frequency is an average, the average of the
frequencies existing at a particular time. If instantaneous frequency is a conditional
average it is natural to then ask for the conditional standard deviation about that
average. Because it is the spread of frequencies at a particular time, we will call it
the instantaneous bandwidth and use Bt or a ,it to denote it. We now discuss a
number of arguments that point to taking
A,(t)
Bt = wlt = A(t) (13.2)
178
Sec. 2 Instantaneous Bandwidth 179
The second term in this expression averages all the deviations of the instantaneous
frequency from the average frequency and certainly corresponds to our intuitive
understanding of spread. But what is the meaning and origin of the first term? In
Section 4.6 we showed that for any joint density, P(x, y), the global spread, Qi,, is
related to the conditional spread, oy x, in the following manner
The similarity of this general result with the bandwidth equation, Eq. (13.3), is sug-
gestive. Associating x with time and y with frequency and comparing, we can im-
mediately infer Eq. (13.2) for the conditional standard deviation. Of course, one
cannot unambiguously equate quantities under the integral sign since we can al-
ways add a quantity that integrates to zero. However, if we added anything that
integrates to zero, the expression for the standard deviation would not remain man-
ifestly positive.
(A" ) 2
AA
z [ 1 + 4 11 (0)] - 2 [ 1 - 40"(0) ] (t) (13.5)
A(t)
As mentioned in Section 9.6, the choice 46"(0) = a leads to a positive spread and in
particular to Eq. (13.2).
There are an infinite number of densities that satisfy the condition 0"(0) = 14
Furthermore, it is not surprising that there are an infinite number of distributions
that could produce a positive standard deviation. What is surprising is that they all
produce the same result. All possible distributions that always give a positive result
for the conditional standard deviation give the same result, namely, IA'(t)/A(t)I.
Purely Frequency Modulated Signals. The above three plausibility derivations all
lead to A'(t)/A(t)I for the instantaneous bandwidth. An important consequence
is that for purely frequency modulated signals, that is, signals whose amplitude is
constant, the instantaneous bandwidth is zero,
Bt = aWit = 0 for s(t) = Ae'`0(t) if A = constant] (13.7)
Physically, what this means is that at each instant of time there is only one fre-
quency, the derivative of the phase. This is in conformity with our intuition that for
a purely frequency modulated signal the instantaneous frequency is known pre-
cisely. Notice also that for this case the global bandwidth is
For purely frequency modulated signals the spread in frequencies comes from the
change in the instantaneous frequency only.
Constant Bandwidth. We now ask for what signals is the instantaneous bandwidth
constant? Solving for IA'(t) /A(t)I = p we immediately have
A(t) - a e-pt (13.9)
Therefore signals whose amplitudes are decaying exponentials have a constant in-
stantaneous bandwidth. For these signals
als the global bandwidth is
Decay Rate Model and Formant Bandwidth.11491 For a decaying exponential the
broadness is proportional to the decay constant. This fact can be used to obtain a
sense of the broadness of an arbitrary signal at time t by fitting an exponential at
that time,
SF(t) = a e-pt cos(wt + coo) (13.11)
Sec. 2 Instantaneous Bandwidth 181
where SF(t) signifies the fitted curve around time t. The adjustable numbers to
obtain the best fit are the three constants a, p, cpo. The fitting is done locally, that
is, we take a small piece of the function at that time and do the best possible fit
by obtaining the constants a, p, and W o. Since we are trying to fit locally, we must
decide on the interval of time around the time of interest. In the case of speech, 5
milliseconds is typically used. Once the curve is fitted, the broadness at that time is
then given by p. For different times we have different sets of "constants".
This idea has been utilized in speech where the p's thus obtained are called for-
mant bandwidth. However, speech is a multicomponent signal and the fitting is
done by a sum of decaying exponentials. In this way we obtain the spread of each
component at each time.
We now want to show the relation between this method of estimating a local
bandwidth and Eq. (13.2). For simplicity we consider one component. Suppose the
signal we are trying to fit is of the form
Expanding the amplitude and phase separately in a Taylor series about some time
t = to, we have
s(t) _ I A(to) + A'(to)(t - to) . . . ] cos( o(to) + cP (to)(t - to) . . .) (13.13)
which gives
Ip A(A'(toto) )
(13.16)
We see that this method of obtaining the spread in frequencies approximates the
instantaneous bandwidth defined by Eq. (13.2).
logs(t) Al (t)
Q(t) = at = + jW,(t) (13.17)
This signal is called the dynamical signal. The relation between the dynamical sig-
nal and the real signal is s'(t) = 3(t) s(t). Now the instantaneous frequency and
instantaneous bandwidth are simply given by
where Im and Re stand for the imaginary and real parts, respectively.
Group Delay and Its Spread. All the concepts and results developed for instan-
taneous frequency can be applied to group delay and its spread. The analogy is
both physical and mathematical. We can think of group delay as the average time
for a particular frequency. It is given by the negative of the derivative of the spec-
tral phase. In complete mathematical analogy with our consideration of the spread
about the instantaneous frequency we have the spread about the group delay,
Bowhead Whale
1.2
0.2
__'p'(t)
0L_
100 150 200 50 100 150 200
FREQUENCY (Hz) FREQUENCY (Hz)
Multicomponent in Time. The above discussion has used situations where the
components have a narrow spread in the frequency direction. Of course, we can
have a situation where components are narrow in the time direction or a combina-
184 Chap. 13 Further Developments
0 2 4 0 0 10 10
RU)
14 to En M 024621012U181820
FII M04cY ("e)
hS
Fig. 13.3 Two examples of multicomponent signals and their energy density spec-
tra. Generally the spectrum cannot be used to determine whether a signal is multi-
component, as in (a). Sometimes it can, as in (b).
lion of the two. The spread along the time dimension of each part must be small
relative to the separation. The separation is the difference in the group delay and
so the condition for a multicomponent signal in time is
Bi(w)
< < 102(w) - V511 (w)I (13.22)
B, (w)
There are many situations where the components may not be well delineated
and this will depend on the amplitudes and phases. As with real mountain ridges
there are many hazy situations in which separate ridges cannot be distinguished.
Given a distribution we know how to obtain the kernel. It is Eq. (9.19). Applying
that formula we have
f ejet+jrW'(t) dt
0(9, T) = f eet+3l w(t+T/2)-w(t-T/2)i (13.24)
dt
This kernel insures that we get Eq. (13.23) for signals that are purely frequency mod-
ulated. Note that it is a functional of the signal.
0(0, -0 =
f ei' ' (t) dt
(13.28)
f ealw(t+2T)-w(t-2r)1 dt
186 Chap. 13 Further Developments
which shows that the marginal cannot be satisfied exactly for this kernel. However,
it is approximately satisfied. To see that, expand W(t+r/2) and cp(t+r/2) in a Taylor
series
cp(t + 2r) - cp(t - 2r) - r cp'(t) (13.29)
and
Oe-7-P'(0
f f
Or=
ei'w'(t) dt
f ei(w(t+2T)-w(t-z1T)l dt
ei'rw'(t) dt
-
f e-7-P'(0 dt =
1 (13.30)
These results are clearly unsatisfactory and argue that there should be another for-
mulation of the general class that takes this difficulty into account in an exact way.
However, that problem is unsolved as of this writing.
Concentration for Bilinear Distributions. In general we can not have a bilinear dis-
tribution satisfying the marginals that is a delta function along the instantaneous
frequency. The reason is that the delta function is positive and we know that we
cannot have positive bilinear distribution satisfying the marginals. The seeming
exception of the Wigner distribution for the signal given by Eq. (8.43) is readily ex-
plained because for that signal the Wigner distribution can be expressed in a form
which is a not bilinear, but of higher order. 1280, 1411 The fact that the bilinear distribu-
tions may go negative presents a problem in the definition of concentration, since
for example standard deviations may go native. Using the square of a distribu-
tion as a measure of concentration, janssen 2761 has considered this problem and
shown that for distributions characterized by kernels of the form 0(9, r) = ell',
the Wigner distribution (a = 0) is generally the most concentrated. However the
question for a general kernel remains open.
and where T and Il are the time and frequency lattice intervals. Gabor proposed
that an arbitrary signal be expanded in the form
s(t) _ Cn,m hn,m (t) hn,m(t) = h(t - rnT) ejnstt n, m = -oo, oo (13.32)
n,m
Sec. 6 Gabor Representation and Time-Frequency Distributions 187
8*(u - 2T)8(u + 2T) _ Cn, m' C,,,,,, hn, m' (u - 2T) hn m(u + 2T) (13.34)
n',m' n,m
where
Cn',m',n,m = 47r2
Jffhi,m(U - 2T) hn,m(u + 2T) 0(0,7-) e-j8'-jrw+jAu du d7- d9
(13.36)
Let us now specialize to the Wigner distribution, in which case the Cn',m';n,m can
be done analytically since the hnm's are Gaussians. If we break up the summation
as
W (t, w)
= E I Cn,m I2Wn,m;n,m(t, w) + E Cn',m' Cn,m Wn',r';n,m(t, w)
n,m n' ,m' fin, rn
(13.37)
then the first summation is manifestly positive since the hnm is a Gaussian. If all
the terms of the second summation are added, then we get the exact Wigner distri-
bution. An effective way to add the terms of the second summation is to first add
nearest neighbor terms, then add second nearest neighbors, and so on. The motive
is to systematically add the relative importance of the cross terms.
188 Chap. 13 Further Developments
where An's and un are constants and functions as yet to be determined. Substituting
this expansion in Eq. (13.38) we have
2
1
8(x) un(x - t) a-3" dx (13.40)
27r
n=1 Yn 27r J
which is a sum of spectrograms with windows un. All this depends on whether
indeed we can decompose r as indicated by Eq. (13.39), which we now consider.
Hermitian Functions. If we have a two dimensional function, K(z, z'), that satisfies
the Hermitian property
K(z, z') = K* (z', z) (13.41)
then solving the integral equation
results in eigenvalues and eigenfunctions, An's and un's, which form a complete
set. For this to be possible the kernel must be square integrable. Such functions are
called Hilbert-Schmidt kernels. The kernel, K, is then expressible in terms of the
eigenfunctions and eigenvalues,
00
K(z, z') = un(z) u* (z') (13.43)
n=1 '\n
Sec. 8 Spectrogram in Terms of Other Distributions 189
Now to specialize to our case. Let z = x - t and z' = x' - t and define
K(z, z') = r(-. (z + z'), z - z') (13.44)
Complex Distributions. If r does not satisfy the Hermitian property then it is pos-
sible to expand it in the following form,
00
r 1 un(x-t)v,,(x' -t) (13.47)
n_1 an
where now the un's and vn's are different complete sets. Such a decomposition is
achieved by a method known as singular value decomposition, which is discussed
in Section 13.9. Substituting into Eq. (13.38) we have
00
C(t, w) = 1 1
l \.
27r
f 8(x) un(x - t) a-'"x dx ] (_L
27r f 3(x) Vn(x - t) a-j"x dx I
Wigner distributions. We shall consider the general case where we expand any dis-
tribution in terms of modified distributions. Recall from Section 9.7 that any two
distributions are related by
Ci(t,W) = (13.49)
Jf 912(t'-t,w'-w)C2(t',w')dt'dw'
with
01(0"r) e? 9t+-7rw dB dT (13.50)
912(1,w) =
12 if
47r 02(0,'r)
If the distributions are real, then 912 (t, w) is Hermitian and
where
C2(n)
(t, W) = Jf rrn(t' - t) 17 (W - w) C2(t', w') dt' dw' (13.54)
C(t, w) _
n=1 Qn
1 un(t) V,*, (W) (13.55)
where the un's and v'n are complete sets obtained by solving the coupled integral
equations
The solution of these equations results in common eigenvalues but two complete
sets of functions, the u's and the v's. This decomposition of a two dimensional
function is called a singular value decomposition. Note that for the case C(t, w) _
C' (w, t) we have Hermiticity, in which case we can use the standard approach dis-
cussed in Section 13.9. Since in our case C(t, w) will be a joint time-frequency dis-
tribution, it is in general not Hermitian.
In the general case we need in principle an infinite number of terms to represent
the distribution exactly. However, in many situations the first four or five terms are
sufficient to achieve a very high accuracy. The concept of decomposing a distribu-
tion in this manner was first done by Marinovich and Eichmanni3631 for the Wigner
distribution.
There are two basic reasons for seeking such a decomposition. First, suppose
the signal is in a noisy environment. If we decompose the distribution and keep
only the first few terms we will reduce the noise significantly because the signal is
well represented by the first few terms but the noise is typically spread out over all
the terms. Therefore, by truncating the series after four or five terms we retain most
of the signal but lose most of the noise.
The second reason is for the purpose of classification. The basic idea is that the
a's contain unique characterizations of the time-frequency structure of a distribu-
tion and may be used for classification. Suppose, for example, we have ten signals
and do a singular value decomposition of each one. This will result in ten sets of
singular values and eigenfunctions, all different from each other. Suppose we keep
just the first five terms. For the eigenvalues we will have ten sets of five numbers.
Now suppose we have an unknown signal, one of the ten but in a noisy environ-
ment, and want to classify it. We do a singular value decomposition and compare
the first five singular values to our sets. Since we are comparing numbers the com-
parison is easy and fast. Any reasonable measure of comparison can be used, such
as the Euclidean distance. The closest distance is used to classify the signal.
13.10 SYNTHESIS
Suppose we want to design a signal that has a time-frequency structure of our
choosing. A way to accomplish this is to form the time-frequency distribution and
then calculate the signal that generates it. If the time-frequency distribution we
have constructed is a legitimate one, that is, a representable one, then the sim-
ple inversion formula, Eq. (9.56), yields the signal. However, in general, the time-
frequency distribution we construct will not be a representable one. For example,
suppose we want a Wigner distribution to have certain time frequency structure.
In general we will not be able to construct a representable Wigner distribution be-
cause we will not be able to properly construct the cross terms. We are simply
not that capable. What is done is to seek a signal that reproduces the given time-
frequency distribution as dose as possible, for example, in the least squares sense.
Obtaining a signal from a distribution is the synthesis problem. It was first con-
sidered by Boudreaux-Bartels and Parks[93, 9si for the Wigner distribution. For the
192 Chap. 13 Further Developments
Choi-Williams distribution, where the cross terms are less of a problem, Jeong and
Williams have devised an effective synthesis scheme. P13
Another important application of the synthesis problem arises in the following
circumstance. Suppose we have a time-frequency distribution of some signal that
is multicomponent, and we want the signal that produces only one of the compo-
nents. What we can do is to erase everything in the time-frequency plane but the
component of interest. Having erased parts of a legitimate distribution, the result-
ing distribution will not usually be a representable one because, for example, we
may have erased cross terms which are needed to have a representable distribu-
tion. Nonetheless, we want the signal that generates that component. The synthe-
sis problem will again have to be done in terms of the best signal that generates a
distribution close to the one we have at hand.
A further reason for the synthesis problem that is if we have a signal with noise
and calculate the time-frequency distribution, the noise will typically be spread
throughout the time-frequency plane. If we literally erase the noise but keep the
signal, the result will be the signal and noise only around the signal in time and fre-
quency. Finding the signal this produces this curtailed distribution will, hopefully,
result in a signal with less noise than the original.
C(t , w ) =
47r_
_
2 N s* (u - 1
2
r) s( u +
2
'r) 0(0 , rr) e-iet-iTW+ieu
du dr d8 (13 . 57)
Kernel Choice. Most of the ideas regarding the choice of kernel carry over to the
random case. For example, if we have a wide-sense stationary process, we expect
the marginal in frequency to be white noise. The condition for that, as Poschl45' has
shown, is that 0(0, 0) = 1, which is the condition to obtain the frequency marginal
for the deterministic case. The issue of which is the best kernel to use for estimating
the properties of the random process has been formulated by Amin,1261 who has
shown that the random case necessitates unique constraints on the kernel. He has
developed a comprehensive theory of kernel choice.
have arisen. We assume that we have discrete samples and that the signal is suffi-
ciently bandlimited so that the Nyquist sampling theorem applies: For a bandlim-
ited signal, the signal can be reconstructed from discrete sampled values if the sam-
pling is done at a sampling rate w, > 2w where is the highest frequency in
the signal.
All distributions have at least one Fourier transform to be done. Historically
there have been a number of methods proposed to calculate integrals involving
sines and cosines, for example the method of Fillon. However, the advent of the
Fast Fourier Transform has over shadowed these methods because indeed it is fast.
The simplest distribution is the spectrogram, Eq. (7.5) One chooses a specific
time, t, calculates the modified signal, s(r) h(r - t), as a function of r, then takes
the Fourier transform with respect to T. This is repeated for each time desired. For
-
the Wigner distribution s*(t 2T) 8(t + 2T) is calculated as a function of r, for a
fixed time t and the Fourier transform taken. That gives us the Wigner distribution
of frequencies at time t. The procedure is repeated for any other time. Note that
it appears because of the factor of one half we must have signal values at points
in between the sampled values. If we do have them then of course there is no
problem. But that requires that we over sample by a factor of two. If we do not,
one can, for example, interpolate to get them. Because of this it has been believed
that to construct the Wigner distribution from discrete samples one must sample
the signal at twice the Nyquist rate; otherwise aliasing will occur. That can not be
the case since in principle having a sampled signal at the Nyquist rate allows us to
construct the signal for any time. Having constructed the signal for an arbitrary time
one can then calculate the continuous Wigner distribution. Therefore we see that in
principle we should not have to over sample. For the Wigner distribution, Poletti11
and Nuttalll405) have shown that the Wigner distribution can be computed from a
signal sampled at the Nyquist rate without interpolation of the sampled values or
reconstitution of the continuous signal.
The same considerations apply to the calculation of any of other bilinear time-
frequency distributions. The basic issue is how to define the discrete version of
bilinear time-frequency distributions so that at the discrete time frequency points
they have the same value as the continuous version, and where the calculation can
be performed directly from the discrete version of the signal, sampled at the Nyquist
rate. In addition, the discrete version should satisfy the same general properties of
the continuous version, that is, the marginals and so forth. This is a fundamen-
tal problem that has recently been solved by Morris and Wu,11871 O'Hair and B. W.
Suter,1410', Jeong and Williams,1286) and Cunningham and Williams.J1671
For the general bilinear case three Fourier transforms have to be performed but
for many kernels one of the integrations can be done analytically resulting in the
form given by Eq. (9.9). This is the case, for example, with the Choi-Williams distri-
bution. For such distributions we have two integrals, only one of which is a Fourier
transform. To calculate such a distribution the inner integral, that is, Eq. (9.10) is ef-
fectively done by the rectangular rule. Subsequently the Fourier transform is taken.
We also mention that decomposition of a time-frequency distribution as a sum
of weighted spectrograms is a very effective calculational method. This is described
Sec. 13 Signal Analysis and Quantum Mechanics 195
in Section 13.7. Also the computational approaches for calculating the positive dis-
tributions are described in Section 14.2 and the calculation method for obtaining
optimum kernels is described in Section 11.5.
O(p) =
a-.ill" dq 'P(q) = O(P) e gplh dp
2rrfi ,! O(q) J
(13.65)
where h is Planck's constant divided by 2ir. Moreover, the probability distribution,
P(q), for finding the particle at a certain position is the absolute square of the wave
function and the probability of finding a particle with a certain momentum, P(p),
is the absolute square of the momentum wave function,
Table 13.1. The relationship between quantum mechanics and signal analysis. The formal mathematical
correspondence is (position, momentum) +-+ (time, frequency). The wave function in quantum mechanics
depends on time but this has no formal correspondence in signal analysis.
Position: Time: t
Momentum: Frequency: w
Time: No correspondence
Expected value of position: (q) = f gIV, (q,t)12dq Mean time: (t) = f t Ie(t)l2 dt
Expected value of momentum: (p) = fpl0(P,t)I2 dp Mean frequency: (w) = f w I S(w) I2 dw
Positive Distributions
Satisfying the Marginals
14.1 INTRODUCTION
Wigner showed that manifestly positive bilinear distributions satisfying the time and
frequency marginals do not exist.[' ] The spectrogram, for example, is man-
ifestly positive but does not satisfy the marginals while the Wigner distribution
satisfies the marginals but is not manifestly positive. Historically there has been
a sense expressed in the literature that manifestly positive distributions satisfying
the marginals cannot exist. But they do. They are obtained by a simple procedure
which ensures that the marginals are satisfied. These distributions are, of course,
not bilinear in the signal.
Take
P(t, w) = I S(w) I2 I s(t) I2 f(u, v) (14.1)
where u, v are functions of t and w
u(t) = f t
I s(t') 12 dt' v(w) = J
00
1 S(w') 12 dw' (14.2)
i f0
I
S2(u, v) du = 1 (14.3)
198
Sec. 2 Positive Distributions 199
It is sufficient to define Q(u, v) only for 0 < u, v < 1. Note that u and v are the
cumulative margins, that is, they are the sum of the densities up to given time and
frequency values.
Positivity. The distributions given by Eq. (14.1) are positive as long as 12 is positive
for the range 0 < u, v < 1. cl may or may not be a functional of the marginals or
signal, but to generate all possible positive joint distributions we have to consider
it a functional of the signal. 12311 Examples of cl's satisfying the above conditions are
easily made up; for example,
il(u,v) = 1 - (nui-1 - 1)(mvm-1 - 1) (14.8)
where n, m are any integers.
Relationship Between Kernels. We have introduced the form given by Eq. (14.1)
because it makes the manifestly positive aspect dear. It is possible to obtain this
form using the general lass developed in Chapter 9. There exists a relationship
between the kernels,11281 but generally it is easier to work directly with Eq. (14.1).
The Wigner Distribution for a Chirp. In Section 8.6 we pointed out that for one
signal and one signal only, the Wigner distribution was manifestly positive. That
case is given by Eq. (8.43). The reason it is positive is the it is not really bilinear It
can be obtained[] from Eq. (14.1).
Weak and Strong Finite Support. For the distributions given by Eq. (14.1), the
factor I S(w)12 1 s(t) 12 appears. Therefore, if the signal is zero at a certain time or
the spectrum is zero at a frequency, the distribution will likewise be zero, assuming
that fl is not infinity at those values. Therefore these positive distributions satisfy
the strong finite support property and hence also weak finite support.
Uncertainty Principle. Since these distributions satisfy the marginals, they satisfy
the uncertainty principle.
200 Chap. 14 Positive Distributions Satisfying the Marginals
Pac(t, w)
= Sac(w) I2 3ac(t) 2 9(uac(t), vec(w)) (14.9)
But
Time and Frequency Shifts. For a signal that is shifted by a time to and frequency
wo the distribution is
I2
= I S(w - wo) I2 I s(t - to) 1z(uah(t), vah(w)) (14.15)
But
P(t,
P(w(t) - 1 3(t) 12
w) = I S(w)
121(u
, v) (14 . 19)
Sec. 3 The Method of Loughlin, Pitton, and Atlas 201
Consequences of Positivity. Our experience in everyday life is, of course, with pos-
itive densities. However, our experience in time-frequency analysis has dealt with
either positive distributions that do not satisfy the marginals or non-manifestly
positive ones that do satisfy the marginals. We now examine the nature of time-
frequency distributions that are both manifestly positive and satisfy the marginals.
First let us note that if the marginal is zero at a certain point in time then the
joint distribution must be zero at that time for all values of frequency; the strong
finite support property is satisfied. This is so because to get the time marginal we
add nonnegative pieces, which cannot add up to zero unless the pieces are zero
themselves. This is no different than saying that if there are no people who are six
feet tall, then there are no people at that height for any weight. Similarly, if the
spectrum is zero for a certain value of frequency, the joint distribution will be zero
for all times at that frequency.
Now consider the issue of instantaneous frequency and positive distributions.
Can we obtain the derivative of the phase as the conditional average,
(w)t = 1
2
y w P(t, w) dw ? (t) (14.20)
s(t) Jo
The answer is no in general, but yes sometimes. The general answer of no comes
about by the following argument of Claasen and Mecklenbrauker.11201 Suppose we
have an analytic signal and therefore the spectrum is zero for negative frequencies.
By the arguments given above the joint distribution will also be zero for negative
frequencies. If the distribution is positive the conditional average will be positive
for an analytic signal. This is a satisfying result. However, we know from Section
2.7 that there are analytic signals that produce an instantaneous frequency, defined
as the derivative of the phase, which may go negative. Hence, for those signals,
Eq. (14.20) cannot be satisfied. But those are precisely the signals for which the
derivative of the phase, as a representation of the concept of instantaneous fre-
quency, apparently makes no sense. Therefore, a meaningful question is whether,
for cases where the derivative of the phase meets our expectations of instantaneous
frequency, positive distributions can satisfy Eq. (14.20). No general results regard-
ing this question are known.
I
IS(t`'I)
(a)
co A
w---+
IS(WW
(a)
(A) ,I FMVINERM
I
(b)
(a)
r
OTr- -
(b)
mtr c)
(d)
(d)
Fig. 14.2 For (A), the signal is a sequence of decaying exponentials, shown in (a). The
frequency is constant. In (b) and (c) are the wide band and narrow band spectrograms.
The manifestly positive distribution that satisfies the marginals is shown in (d). (B) is
the same as (A) except that the frequency is continually increasing linearly. (Courtesy of
P. Loughlin, J. Pitton and L. Atlas.)
15.1 INTRODUCTION
In the previous chapters we developed the theory of joint representations for time
and frequency. We now extend the methods to other physical variables. For a phys-
ical quantity, we want to obtain its density, average value, spread, and other results
as was done for the case of frequency. This is accomplished by expanding the signal
in the representation of the physical quantity in the same sense that the Fourier rep-
resentation is the appropriate representation for the physical quantity "frequency".
The basic idea of representing a signal in another representation is to write the sig-
nal as a linear combination of other functions, the expansion functions. They are
obtained by solving the eigenvalue problem for an operator that represents, or is
associated with, the physical quantity of interest. For example, the complex sinu-
soids are the eigenfunctions of the frequency operator, and the frequency operator
is associated with or corresponds to the physical quantity "frequency". We will use
"a" to signify the physical quantity we are studying and script capital A for the
associated operator.
where u(a, t) are the basis functions and F(a) are the "expansion coefficients" or
the "transform of the signal". As we will prove momentarily, F(a) is given by
204
Sec. 2 Orthogonal Expansion of Signals 205
The basis functions, u(a, t) are always functions of two variables, in this case
time and a. The a's are the numerical values of the physical quantity and they may
be continuous, discrete or both. Furthermore, the range of a's may be infinite or
limited. The integration in Eq. (15.2) implies a specific region of integration. For
example, for frequency the range is all possible values, while for scale we will see
that the range is from zero to infinity. In this section we assume that the variable a
is continuous and at the end of this section we address the discrete case.
The functioh F(a) gives us an indication of how important a particular value of
a is for the signal at hand. If F(a) is relatively large only in a particular region we
can then say that the signal is concentrated at those values of a.
In the eigenvalue problem the operator is given and one seeks to find those func-
tions, the u's, that when operated upon give back the same function multiplied by
a number, in this case a. Generally, there are an infinite number of such functions
each paired with an eigenvalue a. The u(a, t)'s are called the eigenfunctions and
the a's are the eigenvalues. Solution of the eigenvalue problem means solving for
both the u 's and a's. For example, the operator d/ dx, operating on ell returns a e"x,
hence eax is an eigenfunction with eigenvalue a. For this case there are an infinite
number of eigenfunctions because we can take any number for a.
A(f+g)=Af+Ag (15.4)
For example, the operation of differentiation is linear because the derivative of the
sum is the sum of the derivatives. However, the operation of squaring is not because
the square of the sum is not the sum of the squares.
An operator is Hermitian or self adjoint if for any pair of functions, f (t) and
g(t),
fg*(t) A f (t) dt = 11(t) {Ag(t)}* dt (15.5)
To prove that an operator is Hermitian one has to show that Eq. (15.5) does indeed
hold for any pair of functions.
206 Chap. 15 The Representation of Signals
F(a') (15.10)
Possible and Actual Values of the a's. One must be very dear about the following
distinction. The solution of the eigenvalue problem gives us the possible values
that the physical quantity may attain, that is, the a's. However, for a particular
signal the actual values attainable are given by the function F(a), which may or
Sec. 2 Orthogonal Expansion of Signals 207
may not range over all possible values of a. For example, any frequency is possible,
but for a particular signal only certain frequencies may exist. Note that solving the
eigenvalue problem has nothing to do with the signal at hand. The eigenvalue
problem is solved once and for all, but F(a) must be recalculated for each different
signaL
The Usage of "spectrum." The set of eigenvalues obtained by solving the eigen-
value problem is often called the spectrum and the terms continuous and discrete
spectrum are used to communicate whether the d s are continuous or discrete. The
use of the word spectrum for the general case is unfortunate, because spectrum is
associated with frequency which is only one particular case. But even more un-
fortunate, the word spectrum is often used to denote the Fourier transform, S(w).
However, the particular usage is usually clear from the context.
First let us prove that it is Hermitian. For any two functions, f and g, we have, by
integration by parts, that
f 9'(t) dt f(t)dt = f9
0
-00
- f f (t) d
9.(t)
dt (15.12)
Since we want to normalize to a delta function we must take c2 = 1/21r and hence
the normalized frequency eigenfunctions are
(15.17)
where t' are the eigenvalues. This is the equation that led Dirac to invent the Dirac
delta function. The solutions are
u(t, t') = 6(t - t') (15.19)
where t' can be any number. Therefore the eigenvalues are continuous. The eigen-
functions are complete and orthogonal.
Discrete Case. If the eigenvalues are discrete, then the notation a,, is used to denote
them, where n is an integer index. By convention, whenever possible, the eigen-
values are arranged in order of increasing magnitude. The corresponding eigen-
functions are denoted by u,, (t) and the eigenvalue problem is written as
Aun(t) = a,, u,, (t) (15.20)
We emphasize that before we have solved the eigenvalue problem we do not know
whether the solutions will be discrete or continuous or both. As in the continuous
case, the eigenfunctions are orthogonal and complete,
As we will see the commutator of two quantities plays a fundamental role. Some
properties of the commutator are
Forming Hermitian Operators. One often forms a new operator from Hermitian
operators and it is important to know whether the new operator is Hermitian. As-
suming that A and B are Hermitian, the following are readily verified to be Hermi-
210 Chap. 15 The Representation of Signals
tian operators:
cA
An
A+B are Hermitian if A and B are Hermitian and c is real (15.30)
[A,Bl+
[A,B]lj
means that operating with the left hand side on an arbitrary function is the same as
multiplying that function by j. To prove this consider
When will Q(A) thus defined be Hermitian? We already know that An is Hermitian
and therefore c,, An is Hermitian if cn is real. But having the cn's real means that
the function Q(x) is real, hence we conclude that
where T is the unit operator. 'typically we leave out the unit operator because it
is understood to be there where appropriate. For example, we write [ T, W] = j
instead of [ T, W] = jZ.
The inverse of an operator that is the product of two operators is given by
(AB)-1 = B-1A-1 (15.42)
This is so because
The Adjoint of an Operator. The adjoint is another operator, denoted by At, which
forces the equality
J
g`Af dt = f f {Atg}' dt (15.44)
If the adjoint of an operator happens to equal the operator itself, then we see that
Eq. (15.44) becomes the definition of a Hermitian operator, Eq. (15.5), and hence
As can be seen from the definition of adjoint, the adjoint of a constant is the complex
conjugate of the constant,
(AB)t = Bt At (15.47)
A= 2(A+At)+ij(A-At)/j (15.49)
Notice that we have multiplied and divided by j in the second term. The impor-
tance of this decomposition is twofold. First, it breaks up an operator into sym-
metric and antisymmetric parts. Second, it expresses an operator as a sum of a
Hermitian operator plus j times a Hermitian operator. This is the operator analog
of writing a complex number in terms of its real and imaginary part.
A,BI++1[A,B1/j (15.52)
The importance of unitary operators is that they preserve normalization when op-
erating on a function. That is, f (t) and U f (t) have the same normalization if U is
unitary,
f (t) 12
dt = U f (t) 12
dt (15.54)
J J
To see this consider
will be unitary if A is Hermitian. To see this, first note that if A is Hermitian then
Ut = e-jAt (15.57)
This is the case since the middlestep is the definition of adjoint, and the right side
follows because of Eq. (15.53).
15.4 AVERAGES
If the density of a is taken to be I F(a) I2, then the average value of a is
In Section 1.4 we saw that we can calculate frequency averages directly from the
signal, without calculating the Fourier transform. This is a special case of the more
general result whereby the average of g (a) can be calculated directly from the signal,
and the calculation of the transform F(a) can be avoided. This is a fundamental
result of operator theory. In particular,
s' (t) g(A) s(t) dt = f/f F' (a') u' (a', t) g(A)F(a) u(a, t) dt da'da (15.65)
J
Now, sg(A) u(a, t) = g(a) u(a, t), we have
J s' (t) g(A) s(t) dt = f/f F ' (a') u' (a', t) g(a)F(a) u(a, t) dt da'da (15.66)
Averages of Hermitian Operators Are Real. We use both (A) and (a) to signify
the average of a,
(a) = fs*(t)As(t)dt (15.69)
It is important to appreciate that the average defined by Eq. (15.69) will always be
real provided the operator is Hem-titian. The reason is that if the operator is Her-
mitian, then Eq. (15.64) holds and the middle term is real if g is a real function.
W = jd
and therefore
which is a relation we have repeatedly used in the previous chapters to simplify cal-
culations.
Sec. 4 Averages 215
(AB) = (15.73)
z([A,B]+)+.Ij([A,B]lj)
which expresses the average value for the product of two operators in terms of the
real and imaginary parts since both ((AB + BA)) and ([ A, B ] /j) are real.
The Covariance. In Section 1.8 we defined the covariance between time and fre-
quency as the average of t cp(t), where W is the phase of the signal. We now define
the covariance of any two quantities byil
Covab = 1(AB+BA) - (A) (B) (15.74)
A simple proof will now be given. Instead of evaluating (TW + WT) in the time
representation we evaluate it in the spectral representation. Using the same steps as
above,
r
(TW+WT) = (2WT)+j = 2f S'wTSdw+j = -2 J S(w)I2 dw = -2(w 0' )
JJ
(15.80)
Since the two methods of calculation must give the same result we have Eq. (15.79).
216 Chap. 15 The Representation of Signals
CaCb>-Z1([A,B]) (15.81)
1(.4
- (A)) s(t) 12dt (15.84)
J
Similarly for B. Furthermore, the signal that minimizes the uncertainty product is
obtained by solving
(B-(B))s(t) = A (A - (A)) s(t) (15.85)
where
A= ([28])
,a
(15.86)
Note that in general A may be complex since [ A, B ] is not Hermitian. As with the
uncertainty principle for time and frequency, Eqs. (3.3) and (3.4), a more general
result can be proven,
Proof of the Uncertainty Principle. First some preliminaries. For convenience de-
fine
Ao = A- (A) Bo = B- (B) (15.88)
and note that the operators A0 and Bo are Hermitian and that their mean is zero.
Furthermore, it is easy to verify that
and
[Ao,Bo]+ = [A,B]+-2A(B)-2B(A)+2(A)(B) (15.90)
Sec. 5 The Uncertainty Principle for Arbitrary Variables 217
and
([Ao, Bo ]+) = 2 COVab (15.92)
With these preliminaries we can now derive the uncertainty principle for arbi-
trary quantities by considering
r
Oa Q6 = Ao s(t) 1 2 dt x Bo 3(t)
1
2 dt (15.93)
J J
2
= I fs(t)AoBos(t)dt (15.95)
where in going from step (15.93) to step (15.94) we have used the Schwarz inequal-
ity, Eq. (3.10), and in going from (15.94) to (15.95) we have used the Hermiticity
property of the operator. Therefore
and hence
2 2
oa Qb >_ ICOVab+2.7([A0,B0]/7)I2 (15.100)
I([A0,Bo])I2
= Cova6+q (15.101)
= Cov b + 4 I ([ A, B ]) I2 (15.102)
which is Eq. (15.87). Since COVab is positive it can be dropped to obtain the more
standard uncertainty principle, Eq. (15.81).
The minimum uncertainty signal is obtained when the two functions are pro-
portional, \ Aos = Bos and Cov , = 0. That is,
A(A-(A))s=(B-(B))s (15.103)
218 Chap. 15 The Representation of Signals
The constant A is thus far arbitrary. To fix it let us first write the above in terms of
Ao and Bo,
0 = (AoBo)+(BoAo) (15.106)
Multiply the first equation by Ao and take average values; then multiply by Bo and
take average values, to obtain
A as + ab /a = 0 (15.109)
/A =
as - 01 2/_\ ([ Ao, Bo]) = ([ A, B ]) (15.110)
(15.111)
2aa
at aW>zl([T,'J)I=aljI = i (15.112)
Chapter 16
16.1 INTRODUCTION
In this chapter we develop the basic methods to study the density of a single vari-
able. The object is to generalize the ideas and methods that have been developed
for the cases of frequency and time to arbitrary variables.
219
220 Chap. 16 Density of a Single Variable
age and therefore it can be calculated directly from the signal by way of
eyaACnun(t)dt (16.12)
JEECkUk(t)
n k
f n k
Ck uk(t)
a?a"
cn un(t) dt (16.13)
= E IcnI2e'aa (16.15)
n
Sec. 2 Density of a Single Variable 221
which gives
The only values which are not zero are the an's and their density is
P(an) = I Cn I2 (16.19)
and therefore
M(r) = s' (t) s(t + r) dt (16.22)
J
The distribution is given by
I S(w) 12 (16.25)
as expected.
(16.26)
R = W (16.29)
and use the operator methods developed. Solving the eigenvalue problem in the fre-
quency representation
WO w) = ru(r, w)
u(r, (16.30)
gives
u(r, w) _ 6(w - wo/r) _ 6(r - wo/w) (16.31)
Calling G(r) the inverse frequency transform we have
We now obtain an interesting form analogous to the one we obtained for frequency,
Eq. (1.90). Rewrite Eq. (16.34) as
to write
rr 1 / l 1
Since the mean value of a Hermitian operator has to be real we must have
J
('8
s(t)I2dt=O (16.38)
which leaves
W8 _ 11 d e
_ 1 A' (t)
+'P (t) (16.40)
8 a j dt j A(t)
The real part is gyp' (t), and using Eq. (16.39) we immediately have
(w) = (16.41)
As with the average, the standard deviation can also be written directly in terms of
the signal,
Q2
a
= f (a - (a))' I F(a) I2 da (16.42)
As \ 2
1 -(A))s(t) dt (16.44)
2
As
(A dt (16.45)
I
The first two terms are real since (A) is real, hence
QQ = f (1)a(t) 12
dt + f [(k) R - (A) J
2 18(t) 12 dt (16.46)
and therefore
(Ow)2 = J A2
(*) A2(t)dt +
J
(U)))2
A2(t) dt (16.48)
and
(AF)2
Q2 = f 1 F(b) I2 db + f [(f) - (AJ F(b) 12 db (16.53)
at = f (B'(``')2 1B(w)12dw
+J 1S(w)12dw (16.57)
`\ B(w)
Chapter 17
Joint Representations
for Arbitrary Variables
17.1 INTRODUCTION
We now generalize the methods we developed for time-frequency representations
to obtain joint representations for arbitrary quantities. In generalizing we will use
the time-frequency case for examples so that we may check by recovering the known
results. In the next chapter we specialize to the study of scale.
17.2 MARGINALS
For two quantities, a and b, represented by the operators A and B we seek joint
distributions, P(a, b). In the previous chapter we saw that the densities of a and b
are given by I F(a) 12 and I F(b) 12, where F(a) and F(b) are the a and b transforms,
respectively. The joint density should therefore satisfy the marginal conditions
M(a, Q) = (eiaa+_,Ab) =
ff e3cra+j13'P(a,
b)da db (17.3)
225
226 Chap. 17 Joint Representations for Arbitrary Variables
-4
e7aAejOB (17.7)
-., e3RB e3aA (17.8)
Since there are an infinite number of possible orderings, there are an infinite num-
ber of distributions. It is this ordering ambiguity that gives rise to the general class
of time-frequency distributions and we will see that it also gives rise to a general
class for arbitrary variables.
The method for obtaining a distribution is as follows: Choose an ordering, that
is a characteristic function operator, M (a, p), calculate the characteristic function
by way of Eq. (17.5), and then obtain the distribution using
P(a, b) 47r2
ff(M(a, /3)) e-jaa-j)" da d/i (17.10)
Since a and,6 are real numbers and A and B are Hermitian, the quantity aA +08
is also Hermitian. Therefore the solution to the eigenvalue problem gives rise to a
complete set of eigenfunctions, the u's. Any signal can hence be expressed in terms
of them by
s(t) = / u(A, t)F(.1) dA (17.12)
Now consider
&aA+,iABS(t) = ejaA+iPB f u(A, t)F(A) dA = e'-u(.\, t)F(A) dA (17.14)
J
Substituting for F(A) we have
where
G(t, t') = r e&'` u' (A, t') u(A, t) dA (17.16)
This method works quite well in many cases but has the disadvantage that the
eigenvalue problem must be solved.
Method 2. The second method seeks to directly simplify the operator e3cA+jsB.
This, however, is one of the most difficult problems in mathematical analysis and
simplification has thus far been achieved only for special cases. One such case is
when both operators commute with their commutator,
in which case[589]
This relation holds for time and frequency since their commutator is a number and
hence commutes both with the time and frequency operator. We have already used
this method in Chapter 10.
Another special case is when
where
= 1
j.C2
[1-(1+j/9c2)e -jlic2
] (17.29)
This case will arise when we study the scale operator. These relations are proved in
the appendix. Note that for c2 = 0, Eq. (17.24) holds and we have
and it will satisfy the marginals of the original distribution if the kernel satisfies
0(0, Q) = 0(a, 0) = 1 (17.34)
Equation (17.33) is the general class for arbitrary variables. It can be written in the
triple integral form by using Eq. (17.5) for M,
47r
12 ff &1(a, /0)
02(a, /0)
M2 (a,,3 ) a -j-a -jpb da dQ ( 17 . 41 )
230 Chap. 17 Joint Representations for Arbitrary Variables
and since
M2(a,0) = if ejaa+jOl P2 (a, b) da db (17.42)
we have
with
ejaa+j Ob 01 (a, A) do df3
g(a, b) = 12
(17.45)
4-7r J 02(a,,3)
P(b) = -1 f R(f3)
a-jOb d,3 (17.46)
P(a, b) =
2- f R0((J) a-j$b dO (17.47)
as = f (As)8
2 18(t) I2 dt + f [
I 8 R (A 2 18(t) 12 dt (17.50)
Comparing it to the general result that connects the global and local standard de-
viation, Eq. (4.71),
we conclude that the instantaneous value, at, and its conditional standard devia-
tion, aalt, are given by
at=(a )t As lJ 2It As
2
(17.52)
s
va
R
u(a,t)
Because the operator is Hermitian, the eigenvalue, a, is real, and Eq. (17.52) gives
at = a aaIt = 0 (17.54)
We see that in an eigenstate the instantaneous value is equal to the global average
and the local standard deviation is zero for all time.
(A) = f (A )t I s(t) I2 dt =
f()R (17.55)
f s(t)`As(t) dt = f As I S(t) I2 dt
(17.56)
f f (As)
L 8 R
+
j( 3 I.J
I
I s(t) 12 dt (17.57)
f (As) R I
s(t) I2 dt (17.58)
232 Chap. 17 Joint Representations for Arbitrary Variables
where the second term in Eq. (17.57) is zero because we know that the average value
of a Hermitian operator is real.
where F(b) is the signal in the b representation. Comparing with Eq. (17.51) gives
(A(b)F(b)
ab = (a )b = (17.60)
F(b) ) R
z (A(b)F(b)12
)
aalb (17.61)
F(b) r
Again, we expect that the global average be the average of the conditional value,
12
(a) (A )b i F(b) db = 1()R I F(b) I2 db (17.62)
The proof is identical to that given in the previous section where b was time.
For a we take time and for b we take frequency. Hence F(b) is the spectrum S(w)
Writing the spectrum as
S(w) = B(w) e"ti(w) (17.63)
we have
TS
-S11 dd S = 1B'(w)-
B(w)
w
S
= j ( ) (17 . 64)
Therefore
(t)W = -V),(w) (17.65)
which is the group delay. Its standard de viation is, according to Eq. (17.61),
2
at1w ( 17 . 66)
Sec. 10 The Covariance 233
b CA(b)F(b)
F(b) (A)(B) (17.68)
rt
We could have equally well put ourselves in the a representation, in which case
) R \ - (A) (B)
_ (a (B(a)F(a)\ (17.70)
F(a) /
These two definitions are identical because
We prove this by showing first that the definition given in Section 15.5 for the co-
variance is the same as given above, namely
Consider
Ca
F(a))) 2
(a( F(a)))+ 2 (a f a
F()))*) (17.73)
= 2[JF'ABF+JaFBtF'1 (17.74)
= 2(AB+BA) (17.75)
If we start with the expression given by the left hand side of Eq. (17.71), the identical
derivation leads to the same result. Hence we conclude that the two expressions
for the covariance, Eqs. (17.68) and (17.70), are identical.
234 Chap. 17 Joint Representations for Arbitrary Variables
where t is the fixed time of interest, ,r is the running time, and h(t) is the windowing
function. Instead of taking the Fourier transform of this modified signal, we take
the b transform,
Ft (b) = Js(r)h(T_t)u*(b,T)dT (17.77)
In analogy with the short-time Fourier transform this shall be called the short-time
b transform. The joint time - b density is then
2
I2
P(t, b) = I Ft (b) = fs(r)h(r _t)u*(b,T)dT (17.78)
The Short-a b Transform. We first recall that the signals in the a and b representa-
tions, are given by
F(a) = u' (a, t) a(t) dt (17.79)
J
F(b) = u' (b, t) s(t) dt (17.80)
J
These two equations are the transformation from the time representation to the
a and b representations. However, we can transform directly from the a to the b
representation,
F(b) = Jcx(b,a)F(a)da (17.81)
which can be verified by substituting it into Eq. (17.81). Now we window the signal
in the a representation to obtain a modified signal,
where a' is the running variable, and a is the fixed value of interest. Taking the b
transform of the modified signal, we have
and it is reasonable to call this the short-a b transform. The joint density is therefore
r 2
e-jaa-jOl
P(a, b) 412 ffJ us* (u) ejoT+jAwus(u) 0(9, 7-) du da dQ (17.87)
Similarity and Unitary Transformation. To see the relation between the form given
by Eq. (17.87) and the general procedure developed in the previous sections we first
have to define the concept of a similarity and unitary transformation of operators.
If we have an operator and form a new operator, A', according to
A' = U-1 AU [ similarity transformation 1 (17.90)
then that is called a similarity transformation. If, in addition, the operator is unitary,
then U-1 = Ut, in which case
A' = Ut AU [ unitary transformation 1 (17.91)
and the transformation is called a unitary transformation. The reason for defining
a unitary operator transformation in this way is that by doing so we obtain the im-
portant result that averages retain their form in both representations. In particular,
if we call s' the transformed signal
s' = Uts = U-1s (17.92)
then
(a') = fs'(t)..4"s'(t)dt (17.93)
fs*(t)As(t)dt (17.95)
(a) (17.%)
An important property of a unitary transformation is that the eigenvalues of
operators remain the same. Also, Hermiticity is retained; that is, A' is Hermitian
if A is. Another important property of a unitary transformation is that it leaves
algebraic equations functionally the same. In particular, it leaves the commutator
invariant. If
[A,B] = C (17.97)
then
[A!, B'] = C' (17.98)
To see this, multiply Eq. (17.97) on the left by Ut and on the right by U
Ut[A,B]U = UtCU =C' (17.99)
But
Ut [ A, B ]U = Ut (AB - BA)U (17.100)
= [A',B'l (17.103)
Sec. 12 Unitary Transformation 237
To prove this we calculate the characteristic function and distribution for A and B
as per the general procedure given by Eqs. (17.3) and (17.4),
where in going from step (17.107) to (17.108) we have used the fact that
e.70g
e3a'A = Ut e&aT U ; = Ut e3,31V U (17.109)
which can be readily proven. The distribution is the double Fourier transform of
M, which gives Eq. (17.87).
Therefore for the case where the two new variables are connected to time and
frequency by a similarity transformation, the general class of distributions for a and
b is obtained from the general class of time-frequency distributions by substituting
the transformed signal for s(t). This simplification can be used whenever applica-
ble.
Then we do not have to recalculate their general class. It can be obtained from Eq.
(17.110) by substituting the transformed signal for s(t),
238 Chap. 17 Joint Representations for Arbitrary Variables
The proof is the same as in the time-frequency case. This shows that variables con-
nected by a unitary transformation have distributions that are functionally identi-
cal.
Method 1. The straightforward way to obtain such a distribution is to use the gen-
eral class of time-frequency distributions and the method of Section 4.7 to transform
distributions. The distribution in time and inverse frequency, P(t, r), is given by
we obtain
P(t, r) = r2 P(t, wo/r) (17.115)
where the P on the right hand side is the time-frequency distribution. Any time-
frequency distribution can be used for P(t, w). If we use the Wigner distribution
P(t, r) - rwe
W(t, wo/r) (17.116)
1 WO /
s'(t- 2r)e-21 ?O/'s(t+ zT)dTr (17.117)
27r r2 J
To obtain the general class of time-inverse frequency distributions we can use the
general class of time-frequency distributions, Eq. (9.1), together with Eq. (17.115).
Doing so yields
Method 2. Alternatively, although in this case much harder, we can use operator
methods. Here we do not need to, since we are mating a transformation of ordinary
functions; nevertheless, it is of some interest to do so. Take
The solution is
u(A w) = 1 e-71J1W-WORIn IWI1/B
(17.125)
21x9
and therefore
&BT+7RTtS(w) = r S1""/ 5(9 + w' - w) S(w') dw' (17.127)
SON
S(w - 0) exp I j In W w e1
(17.128)
W+ 2B
JS*(W+0)S(w_9)exp[jn]dw
1
(17.130)
1
240 Chap 17 Joint Representations for Arbitrary Variables
i
P(t, r) fffS*(w+9) S(w - 10) exp I j wOK In w + i e I e-yet-j"'' dO dw
z
(17.131)
17.14 APPENDIX
The simplification of eA+B is generally difficult but for the case where'
The idea is to obtain a differential equation for f (A) and solve it. The solution,
hopefully, will be a simpler form than what we started out with.
Before proceeding we establish some operator relations. Taking the commutator
of A with [ A, B ] we have
[A,[A,B]]=[A,aA]=0 (17.135)
Also,
[ B, [ A, B ] ] = a[ B, A]= -a/- a 2 A (17.136)
There is a well known relationIM91 that is very useful for simplification of operator
identities,
,\2 V
e'\ASe -\A=B+.1[A,B]+ 2i[A,[A,B]]+ 3i[A,[A,[A,B]]]+... (17.137)
1 For the sake of neatness we use here the notation [ A, B ] = 0 + aA rather than [ A, B ] = cl + c2A
used in the text. The a, 0 used in this appendix should not be confused with their usage in the text.
Sec. 14 Appendix 241
Solving, we have
f (A) = e-a9/a
a-aA {1 - (1 + \a) a-a} (17 . 143)
as
Hence
e AB e AA e -a(A+B) = e - aA/d a - iAA (17.144)
or
ea(A+B) = eap/a ea.4 eaB e.1A (17.145)
Taking.1= 1 gives
Scale
18.1 INTRODUCTION
In this chapter we develop, using the general methods of the previous chapters,
the basic properties of scale. Scale is considered to be a physical attribute just like
frequency. The frequency transform, the Fourier transform, allows us to ascertain
the frequency content of a signal. Similarly, we will develop the scale transform,
which allows the determination of the scale content. The first objective is to ob-
tain the operator that represents scale. The scale transform is obtained by solving
the eigenvalue problem for the operator. Having the scale operator and transform
allows us to obtain average scale, scale bandwidth, instantaneous scale, and other
properties in complete analogy to frequency. Scaling can be applied to any variable.
For concreteness we develop scale for time functions and subsequently consider the
scaling of other variables, such as frequency. In the next chapter we develop joint
representations of scale and time and frequency.
C = 2(TW+WT) = (18.1)
That this operator gives results that conform to our intuitive sense of scaling will be
seen as we develop its properties. In the time representation C is given by
(18.2)
Sec. 2 The Scale and Compression Operator 243
Using the commutation relation for time and frequency, (TW - WT) = j, it can
be written in the following alternative forms,
C=TW-2j=WT+2j (18.3)
We expect a similar result for the scale operator in that it will define a compression
operator. In fact,
s(t) = eal2 s(ea/2t) (18.5)
That is, eiac compresses functions with the factor ea/2. If we take In a for a we
have
ejInaC 8(t) = (18.6)
s(at)
The significance of the factors, ea/2 or f, is that they preserve normalization. They
entered in an automatic way because the operator ejc is unitary.
Equation. (18.6) is basic and we give two proofs. The first proof relies strictly on
the algebraic properties of the operator; the second depends on the eigenfunctions
of scale and is presented after we obtain them. To show Eq. (18.6) consider first the
action of C on tn,
Now
ejaCtn = `` Grktn = 00 q (-j)k(n + 2)ktn
= eQ(n+1/2)tn (18.10)
k. k. l
k=O k =O
To obtain the action of ejac on an arbitrary function, s(t) , we expand the function
in a power series
00
s(t) = e&ac E ante = ea/2 E an eantn = ea/2s(eat) (18.11)
n=0 n=0
244 Chap. 18 Scale
Basic Commutation Relations. Unlike the commutator for time and frequency the
commutator of scale with time or frequency does not produce a number. The com-
mutation relations are
All these relations are important for the simplification of expressions involving the
scale operator.
_jtdy(c,t) _-
1
?
y(c, t) = c y(c, t) (18.16)
dt
The solutions, normalized to a delta function, are
1 ejctnt
y(c, t) = t>0 (18.17)
2?r f
It is possible to obtain this solution only if time is positive, and thus for scaling we
must have an origin in time. This is reasonable since to scale is to enlarge or reduce.
That means we must multiply time t by a number from zero to one for enlargement
and a number bigger than one for reduction. At no time do we multiply by a nega-
tive number. Also, the appearance of the logarithm is reasonable because it has the
effect of putting on an equal footing the range of zero to one (enlargement) with
the range of one to infinity (reduction).
Sec. 4 The Scale Transform 245
Properties of the Scale Eigenfunctions. In Table 18.1 we list the main algebraic
properties of the eigenfunctions and contrast them to the frequency eigenfunctions,
1 e2Wt
u(w, t) = (18.20)
=
2
Notice that with respect to the time variable, multiplication/division in the scale
case corresponds to addition/subtraction in the frequency case.
7(c, tt') = 7(c, t) 7(c' t') u(w, t + t') = u(w, t) u(w, t')
7(c, t/t') = 7(c, t) y' (c, t') u(w, t - t') = u(w, t) u* (w, t')
1 ejctnt
dc (18.22)
2n J D(c) vt-
246 Chap. 18 Scale
1 roo a-jclnt
= s(t) dt (18.24)
27r J VIt
We call D(c) the scale transform. It can also be written as
f s(t) t-jc-1/2 dt
D(c) = 1
2Tr f (18.25)
which shows that it is the Mellin transform of the signal with the complex argument
-jc+ 2 . We know from the last chapter that the density of a quantity is the absolute
square of its transform and therefore the density of scale is
P(c) = I D(c) 12 [Intensity or density of scale] (18.26)
Therefore
I2 D(c) 1 2 (18.30)
Dac(c) =
This a basic result which shows that the scaling of any time function leaves the
energy density scale spectrum unchanged.
Alternative Proof of the Compression Operator. Using the fact that the scale eigen-
functions satisfy e3'5(c, t) = erl2ry(c, at), we have
The Scale Transform of the Sum of Scaled Functions. Let us first consider a signal
composed of the sum of translated functions,
The Fourier transform of s(t) is the sum of the Fourier transform of each term. But
the Fourier transform of f (t - tk) is e&"tMF(w), where F is the Fourier transform of
f (t). Therefore we have
n n
S(w) = E
k=1
F(w)
k=1
ckey"tk (18.34)
The scale transform of each term is ck eic In all E(c), where E(c) is the scale transform
of f (t). Therefore
n n
D(c) = [ Ck a clnakE(c) = E(c) Ck ejclnak
(18.36)
l L=1
IC=1 k=1
If we consider the situation where the amplitudes are arbitrary
n
3(t) = Akf (akt) (18.37)
k=1
then
D(C) = E(C) k C7clnak (18.38)
k=1
Relation with the Fourier Transform. An interesting viewpoint of the scale trans-
form is obtained by considering a new function constructed by replacing time with
the logarithm of time,
ft (t) _ s(lnt) (18.39)
The factor 1/f is inserted to preserve normalization. Now consider the scale trans-
form of ft,
DI (c)
= 1
Jo =
ft(t)
f
ajnt dt (18.40)
248
1 roo a-jclnt
= s(lnt) t dt
27r J
=
1
27r f
oo
s(r) a-jcr dT
Inversely we can consider the scale transform to be the Fourier transform of the
function s(et) et/2. If we define
1 0o a-jclnt
_ s(t) dt
27r o v (18.46)
= D(c) (18.47)
Example 18.1: The Highest and the Lowest Scale Content Signals.
What functions have the lowest or flattest scale content? We approach this by asking
what functions have the lowest frequency content. They are functions that have a flat
Sec. 6 Scale Characteristic Function 249
spectrum. Signals that give rise to flat spectra are impulses, 6(t - to). Therefore, using
our argument above, the functions that have the lowest scale content are
_ 1 e-jelntp
(18.49)
D(c) 2a to
where for the last step in Eq. (18.50) we have used Eq. (18.93). We see therefore that
impulses have the lowest frequency and scale content.
Q = In(y/x) t = xy dx
; ; t do, dt = dx dy ; do dt = yy (18.58)
we have
P(c) =
1
2I Jo
O
f
8' (x) s(y)
e-jc(lny-Inx)
xy
dxdy (18.59)
= 1 D(c) 12 (18.60)
250 Chap. 18 Scale
(c) = (18.62)
J
(CS)
(18.63)
= f7 R I s(t) 12 dt (18.64)
All three formulas will lead to the same result, but the last is the easiest to use.
Calculating
C9s \
= A() +2J (18.65)
1 (t
we see that the real part is t cp'(t) and therefore
The Scale Bandwidth. Similarly, the scale bandwidth can also be calculated by
using any one of the three alternative formulas,
UC = f(c_(c))2ID(c)2dc (18.67)
(Cs12
s(t) 12 dt + r 1 (Cs
\ - (C 12 dt (18.69)
] 2 1 s(t)
S I J 8 R
Sec. 8 Instantaneous Scale 251
where, the last equation is obtained from Eq. (16.46). Again, the last equation is
easiest since we have just calculated the real and imaginary parts of Cs/E, Eq. (18.65).
We immediately have
T tl --2I([T,C])I (18.73)
From Eq. (18.12) we know that [T, C ] = jT and therefore we have the following
uncertainty principle for time and scale,[' 1-111
or 01 2 > (18.74)
ZI(t)I
where (t) is the mean time.
Minimum Uncertainty Product Signal. Using Eq. (15.103), the minimum uncer-
tainty product signal is
where
([T,C)) _ (t) (18.76)
2vt tat
and the solution isE1511
8(t) = k t0' (18.77)
- (w+4) (18.79)
Everything we have derived for time scaling can beimmediately transliterated for
frequency by substituting w for t appropriately. In particular, the frequency scale
eigenfunctions are
1 ejclil W
7W (c, w} = (18.80)
27r V'_
and the transformations between the frequency domain and the scale domain are
G'_9C}I1W
1
S(w) =
2-7r
JD(c) do , w>0 (18.81)
W
e
D,, dw (18.82)
(c) j oy S(w) v -
Z-r
We have used DW (c) to emphasize that we are defining a different transform, the
frequency scale transform. For the frequency scaling representation only the posi-
tive half of the spectrum is considered, which is equivalent to considering analytic
signals. For frequency functions we have
e2c, S(w) = eo/2S(e'w) ej Inoc., S(w) = / S(Qw) (18.83)
f(w
oc = f (tW(W) +2)2B2(w)dw+ '(w)-(c))2B2(w)dw (18.85)
/
where B(w) and ti(w) are the spectral amplitude and phase. Similarly, instanta-
neous scale in the frequency domain is given by
c, = wV (w) (18.86)
Other Domains. For an arbitrary domain, say the a domain, the scale operator is
Ca
2j
(a+a) (18.87)
To convert the equations we have derived for time scaling to scaling in the a domain,
all we have to do is substitute a for t and s(a) for s(t) in any of the formulas. The
above application for frequency was an example of this.
18.11 APPENDIX
We prove the completeness relations Eq. (18.18) and (18.19). Consider the left hand
side of Eq. (18.18)
00 [00 e-jdlnt ejclnt
f0
y* (c', t) y(c, t) dt
1
27r
1
o f f
ao e3(c-c )Int
(18.89)
dt (18 . 90)
27r 10 t
f
00
1 e&(c-c')x dx (18.91)
27r
where in going from Eq. (18.89) to Eq. (18.91) we have made the transformation
x = In t. To prove Eq. (18.19) we first establish the following delta function identity
which is the same answer that would be obtained if we integrate the right hand
side of Eq. (18.93) with f (x). Now consider
1 e-jcut' e3clnt
ry'(c,t')y(c,t)dc (18.95)
J 27r t' V
t b(ln t - In t') (18.96)
1
b(ln(t/t')) (18.97)
19.1 INTRODUCTION
In this chapter we obtain joint representations where scale is one of the variables
and the other variable is time or frequency. We also consider joint representations
of the three variables, time, frequency, and scale. The general approach we use is
the characteristic function operator method developed in Chapters 10 and 17.
= = (&et+)oc)
M(9, Q)
ff SBt+iac P(t, c) dt dc (19.1)
= 4'
P(t' c)
ff M(9, a) e-jet-j` dO do (19.2)
Marginals. The energy densities of time and scale are I s(t) 12 and I D(c) 12, respec-
tively. Therefore, we would ideally like
However, as in the time-frequency case, there are joint representations that do not
satisfy the marginals exactly but are nonetheless useful.
255
256 Chap. 19 Joint Scale Representations
Joint Characteristic Function. For the characteristic function operator we use the
notation M(8, a; T, C). The average of the characteristic function operator is the
characteristic function,
As we discussed in Chapter 17, there are many possible orderings. In Table 19.1 we
list some of the possibilities and the corresponding characteristic functions and joint
distributions. In the appendix we work out the first two choices in detail because
they utilize and illustrate the main mathematical techniques. Marinovich,1179,
Bertrand and Bertrandl"8, 711 and Altes112) introduced the concept of joint scale repre-
sentations. They obtained, by other methods, the first two representations of Table
19.1. Further development of the theory has been made by Rioul,(4Ml Posch,14541 Ri-
oul and Flandrin,14901 Papandreou, Hlawatsch and Boudreaux-Bartels,1421, 42a) Bara-
niuk and Jones, 149,511 Shenoy and Parks,151o, 5111 and Cohen.1141,1511
It is straightforward to verify, by direct integration, that the time-scale distri-
butions of Table 19.1 satisfy the marginals of time and scale, Eqs. (19.3) and (19.4).
However, the easiest way to ascertain that is by noting that each of those character-
istic functions satisfy the characteristic functions of the marginals. That is,
ejOT
M(O,O;T,C) = ; M(o o;? C) = e2'c (19.6)
Of course, checking the characteristic function does not check whether the algebra
to obtain the distribution was correctly done.
Suppose we choose the characteristic function given by ordering one of Table 19.1.
Then
h
--7
1
ejC/2ej8TejC/2
f o
"O e
,)et
s'
(e-/2t)8(e/2t) dt
2 Ja*(e/2t)eaca(etT/2t)dc
f 00
2 ejOT+jC e2jOte,nh(/2)/ e-jc
21r 2 sinh(a/2)
at at
a* (E /2t)a(e/2t) dt a' (e-O/2
2 sinh(a/2)
} s(e/2
i(&/2) ) da
3 exp[jet cosh(a/2)] 1a J/
2
1
cosh(o/2)
e-1 C
J0
)e(e/2
s' (e-/2t)8(e/2t) dt s' (e-O/2 da
cosh(a/2) cosh(a/2))
Hence
Mj(e Q) = 01(0, a)
M2(9,o) (19.11)
02(0, 0')
By taking the Fourier transform of both sides we obtain
with
eet+jc 2(010") d8 du
g(t, c) = 412 JJ
if (19.13)
258 Chap. 19 Joint Scale Representations
with
Rt(CT) if 0(0, a) M(O, Q) e-iet dO (19.15)
= 27r
We shall call Rt the generalized local scale autocorrelation function in analogy with
the local autocorrelation function for frequency, Eq. (9.7). Using Eq. (19.9) for the
general class, Rt (v) is explicitly given by
We use the notation P(t, w, c) to signify the joint distribution and M(9, T, v) to sig-
nify the characteristic function. In Chapter 4 we discussed multivariate joint dis-
tributions but used the two dimensional case for the sake of clarity. We are now
dealing with a three dimensional situation in which case the characteristic function
and distribution are related by
which evaluates to
()5
ffff 0(0, T,a) 3'(e-'/'(4u - 2T)) s(e/2(u+ 2T))
X e-je(t-u)-jrw-jc dO da dT du (19.33)
The marginals of this general class are the general lasses of the marginals. For ex-
ample, consider integrating out scale. Integrating over c gives a 27r6(a) and there-
fore
1
4(9, T, 0) s' (u - r) s(u + z r) a -je(t-u)-jrw dB dr du (19.35)
47x2 JfJ 2
This is precisely the general class of time-frequency distributions, Eq. (9.1), with the
general kernel (B, Or, 0).
19.6 APPENDIX
We derive the distributions corresponding to orderings 1 and 2 of Table 19.1.
The distribution is
a-j`
s`(e`/2t')6(t - t')s(el2t') dv dt' (19.43)
2- JJ
Hence
s(e/2t) do
P(t, C) =
2 J s*(e-l2t) e- (19.44)
with
77 _-1 {1-(1-Q)e}
01
(19.47)
The distribution is
which reduces to
P t c= 27r 2 sinh(Q/2)
as' (e_122 sinh(U/2)) 8 (e-/2 Qt
2 sinh(t7/2)
(19.52)
Bibliography
[1] R. M. S. S. Abeysekera and B. Boashash, "Time frequency domain features of ECG signals: Their
application in P wave detection using the cross Wigner-Ville distribution," Proc. IEEE ICASSP- 89,
pp. 1524-1527,1989.
[2] S. Abrahamsson, B. Brusmark, C. G. Gaunaurd and H. C. Strifors, "'Target identification by means
of impulse radar," Proc. SPIE, vol. 1471, pp. 130-141,1991.
[31 S. Abrahamsson, B. Brusmark, H. C. Strifors, and C. G. Gaunaurd, "Extraction of target features
in the combined time-frequency domain by means of impulse radar," Proc. SHE, vol. 1700, pp.
102-114,1992.
[4] M. H. Ackroyd, "Instantaneous and time-varying spectra - An introduction," Radio Electron. Eng.,
vol. 239, pp. 145-152,1970.
[5] Ackroyd, M. H., "Short-time spectra and time-frequency energy distributions," J. Acoust. Soc. Am.,
vol. 50, pp. 1229-1231,1970.
[6] Adamopoulos, P G. and Hammond, J. K, "The use of the Wigner-Ville spectrum as a method of
identifying/characterizing nonlinearities in systems," Proc. IEEE ICASSP- 87, vol. 3, pp. 1541-1544,
1987.
[7] Adamopoulos, P G., Hammond, J. K, and Lee, J. S., "The Wigner distribution, multi-equilibria
nonlinear systems and chaotic behavior," Proc. IEEE ICASSP- 88, pp. 2216-2219,1988.
[8] D. V Ageyev, "The active band of the frequency spectrum of a function of time," Tiudy Gor7cooskogo
Politekhnicheskogo In-ta, vol. 11, pp. 5,1955.
[9] J. B. Allen and L. R. Rabiner, "A Unified Approach to STIT Analysis and Synthesis," Proceedings of
the IEEE, vol. 65, pp. 1558-1564,1977.
1101 J. B. Allen, "Short-Time Spectral Analysis, Synthesis and Modification by Discrete Fourier Trans-
form." IEEE Trans. Acoust., Speech, Signal Processing, vol. 25, pp. 235-238,1977.
[11] R. A. Altes, "Detection, estimation and classification with spectrograms," J. Acoust. Soc. Am., vol.
67, pp. 1232-1246,1980.
[12] R. A. Altes, "Wideband, proportional-bandwidth Wigner-Ville analysis," IEEE Trans. Acoust.,
Speech, Signal Processing, vol. 38, pp. 1005-1012,1990.
[13] P O. Amblard and J. L. Lacoume, "Construction of Fourth Order Cohen's Class: A Deductive
Approach," Proc. IEEE-SP Int. Sympos. Time-Frequency and T ne-Scale Analysis, pp. 257-260,1992.
[14] M. G. Amin, "The application of the GLPW technique to time-varying spectrum estimation," in:
Applied Signal Processing, ed. M. H. Hamza, pp. 298-302, Acta Press, 1985.
[15] M. G. Amin, "Tune-varying spectrum estimation of a general class of nonstationary processes,"
Proceedings of the IEEE, vol. 74, pp. 1800-1802,1986.
[16] M. G. Amin, "On the application of the single pole filter in recursive power spectrum estimation,"
Proceedings of the IEEE, vol. 75, pp. 729- 731,1987.
[171 M. G. Amin, "Tune and lag window selection in Wigner-Ville distribution," Proc. IEEE ICASSP- 87,
vol. 3, pp. 1529-1532,1987.
[18] M. G. Amin, "Computationally lag invariant recursive spectrum estimators," IEEE Trans. Acoust,
Speech, Signal Processing, vol. 35, pp. 1713-17Z4,1987.
263
264 Bibliography
[19] M. G. Amin, "Recursion in Wigner distributions ," Proc. SPIE, vol. 975, pp. 221-231,1988
[20] M. G Amm, "Mixed time-based spectrum estimators," IEEE Trans Acoust., Speech, Signal Process-
ing, vol. 36, pp. 739-750,1988.
[21] M G. Amm and M. T. Shiavoni, "Time-varying spectrum estimation via multidimensional filter
representation," Proc SPIE, vol. 1152, pp. 437-448,1989.
[22] M. G. Amin and E Allen, "Separation of signals and cross-terms in Wigner Distribution using
adaptive filtering," Proceedings of the 32nd Midwest Symposium on Circuits and Systems, vol. 2, pp.
853-856,1989
[23] M. G. Amin and F Allen, "A novel approach to cross-terms elimination in Wigner distribution
using adaptive filtering in frequency domain," Proc SPIE, vol. 1152, pp. 496-505,1989.
[24] M. G. Amin, "High temporal resolution estimators through reduced rank periodograms," Proc.
IEEE ICASSP- 90, vol. 4, pp. 2471-2474,1990.
[25] M G. Amin, "Statistical Trade Offs in Modern Tune-Frequency Kernel Design,' Proc of Twenty-
Fifth Annual Asilomar Conference on Signals, Systems and Computers, pp. 359-363,1991.
[26] M. C. Amin, "Time-frequency spectrum analysis and estimation for non-stationary random pro-
cesses," in: Time-Frequency Signal Analysis, edited by B. Boashash, pp. 208-232, Longman Cheshire,
1992.
[27] M. G. Amin, "Spectral Decomposition of time-frequency distribution kernels," Proceedings of the
Asilomar Conference on Circuits, Systems, and Computers, vol. 1, pp. 488-492,1992.
[28] M. G. Amin and W. J Williams. "High Spectral Resolution Time-Frequency Distribution Kernels,"
Proc. SPIE, vol. 1770, pp. 302-314,1992.
[29] M. G. Amin, P. Davis, and F. Allen "An Application of the LMS Algorithm in smoothing Pseudo
Wigner Distribution," IEEE Transactions on Signal Processing, vol. 41, pp. 930-934,1993.
[30] M. G. Amin and W. J. Williams, "Deterministic exponential modeling techniques for high spectral
resolution time-frequency distributions," Proc. IEEE ICASSP- 93, vol. 3, pp. 253-256,1993.
[31] M. G. Amin and J. E Carroll, "A linearly constrained minimization approach for time-frequency
kernel design," Proc. 27th Asilomar Conf. on Sigs., Syst. and Camps., pp. 476-480,1993.
[32] M. G. Amin, "Introducing the spectral diversity," IEEE Transactions on Signal Processing, vol. 41, pp.
185-193,1993
[33] M. G. Amin, "The running time-frequency distributions," Journal of Circuits, Systems, and Signal
Processing, to appear, 1995.
[34] M. G. Amin, "Spectral decomposition of the time-frequency distribution kernels,' IEEE Transac-
tions on Signal Processing vol. 42, pp. 1156-1166,1994.
[35] J. C. Andrieux, M. R. Feix, G. Mourgues, P. Bertrand, B. lzrar, and V. I Nguyen, "Optimum
smoothing of the Wigner-Ville distribution," IEEE Trans. Acoust., Speech, Signal Processing, vol. 35,
pp. 764-769,1987.
[36] R. A. Athale and J. N. Lee and E. L. Robinson and H. H. Szu, Acousto-optic processors for real-
time generation of time-frequency representations," Opt. Lett., vol. 8, pp. 166-168,1983
[37] L E Atlas, Y. Zhao, R. J. Marks II, Application of the generalized time-frequency representation
to speech signal analysis," IEEE Proc. Pacific Rim Conf on Comm., Comp. and Sig. Proc., pp. 517-520,
1987.
[38] L. E. Atlas and W. Kooiman and P. J Loughlm and R. Cole, "New non-stationary techniques for
the analysis and display of speech transients," Proc. IEEE ICASSP- 90, pp. 385-388,1990.
[39] L. E. Atlas, P Loughlin, J. Pitton and W. Fox, 'Applications of cone-shaped kernel time-frequency
representations to speech and sonar analysis," 2nd Int. Symp. on Sig. Proc. and its Applications, pp.
882-885,1990.
(40] L. E. Atlas and P J. Loughlin and J. W. Pitton, "Signal analysis with cone kernel time-frequency
representations and their application to speech," in. Time-Frequency Signal Analysis, edited by B.
Boashash, pp. 375-388, Longman Cheshire, 1992
[41] L. E. Atlas, J. Fang, P Loughlin, and W. Music, "Resolution advantages of quadratic signal process-
ing," Proc. SPIE, vol. 1566, pp. 134-143,1991.
Bibliography 265
[42] L. E. Atlas, P Loughlin, J. Pitton, "Truly nonstationary techniques for the analysis and display of
voiced speech," Proc. IEEE ICASSP- 91, pp. 433-436,1991.
[43] L. E. Atlas and J. Fang, "Quadratic Detectors for General Nonlinear Analysis of Speech," Proc. IEEE
ICASSP- 92, pp. 11-9-11-12,1992.
[44] R. Bamler and H. Glunder, "The Wigner distribution function of two-dimensional signals; coher-
ent optical generation and display," Opt. Acta., vol. 30, pp. 1789-1803,1983.
[45] R. G. Baraniuk and D. L. Jones, "Optimal kernels for time-frequency analysis," Proc. SPIE, vol.
1348, pp. 181-187,1990.
[46] R. G. Baraniuk, D. L. Jones, T Brotherton, and S. L. Marple, "Applications of adaptive time-
frequency representations to underwater acoustic signal processing," in: Proceedings of the 25th
Asilomar Conference on Signals, Systems, and Computers, vol. 2, pp. 1109-1113, 1991.
[471 R. G. Baraniuk and D. L. Jones, "Signal-dependent time-frequency analysis using a radially Gaus-
sian kernel," Signal Processing, vol. 32, pp. 263-284,1993.
[48] R. G. Baraniuk and D. L. Jones, "A signal-dependent time-frequency representation: optimal
kernel design," IEEE Trans. on Signal Processing, vol. 41, pp. 1589-1602, 1993.
[49] R. G. Baraniuk and D. L. Jones, "Warped wavelet bases: unitary equivalence and signal process-
ing," Proc. IEEE ICASSP- 93, pp. 320-323,1993.
[50] R. G. Baraniuk, "A signal transform covariant to scale changes," Electronics Letters, vol. 29, pp.
1675-1676,17,1993.
[511 R. G. Baraniuk and D. L. Jones, "A signal-dependent time-frequency representation: Fast algo-
rithm for optimal kernel design," IEEE Trans. Sig. Proc., vol. 42, pp. 134-146,1994.
[52] it G. Baraniuk and L. Cohen, "On Joint Distributions for Arbitrary Variables," IEEE Signal Process-
ing Letters, submitted, 1994.
[53] R. G. Baraniuk and D. L. Jones, "Unitary equivalence: a new twist on signal processing," IEEE
Trans. on Signal Processing, submitted, 1993.
[54] N. E Barber and F UrseIl, The response of a resonant system to a gliding tone," Phil. Mag., vol.
39, pp. 345-361,1948.
[55] D. T Barry, N. M. Cole and J. C. Wood, "Time-frequency analysis of resonating biological systems,"
in: ISSPA 90 - Signal Processing, vol. 2, pp. 930-933,1990.
[56] D. T Barry, N. M. Cole and J. C. Wood, "Muscle sounds occur at the resonant frequency of skeletal
muscle," IEEE Blamed. Engr., vol. 37, pp. 525-531,1990.
[57] D. T Barry and J. C. Wood, "Time-frequency analysis of the human first heart sound," Proceedings
of the IEEE Society of Engineering in Medicine and Biology, pp. 2097-2098,1991.
[58] D.1 Barry, "Fast Calculation of the Choi-Williams Distribution," IEEE Trans. on Signal Processing,
vol. 40, pp. 450-455,1992.
[59] H. 0. Bartelt, K H. Brenner and W Lohmann, "The Wigner distribution function and its optical
production," Opt. Commun., vol. 32, pp. 32-38, 1980.
[60] M. J. Bastiaans, "The Wigner distribution applied to optical signals and systems," Opt. Commun.,
vol. 25, pp. 26-30,1978.
[61] M. J. Bastiaans, "Wigner distribution function and its application to first order optics" Jour. Opt.
Soc. Am., vol. 69, pp. 1710-1716,1979.
[62] M. J. Bastiaans, "Signal description by means of a local frequency spectrum," Proc. SPIE, vol. 373,
pp. 49-62,1981.
[63) M. J. Bastiaans, "On the sliding window representation in digital signal processing," IEEE Trans.
Acoust., Speech, Signal Processing, vol. 33, pp. 868-873,1985.
[64] E. Bedrosian, "A product theorem for Hilbert transforms," Proceedings of the IEEE, vol. 51, pp. 868-
869,1%3.
[65] S. C. Bekir, "A contribution to the unaliased discrete-time Wigner distribution," J. Acoust. Soc Am.,
vol. 93, pp. 363-371,1993.
[66] L. It Berriel-Valdos, C. Gonzalo and J. Bescos, "Computation of the Wigner distribution function
by the Hartley transform," Opt. Commun., vol. 68, pp. 339-344,1988.
266 Bibliography
[671 M. V. Berry, "Semiclassical Mechanics in Phase Space: A Study of Wigner's Function," Phil. Trans.
Roy. Soc., vol. A-287, pp. 237-271,1977.
[68] J. Bertrand and P Bertrand "Representation temps-frequence des signoux," C. R. Acad. Sc., vol.
299 (serie 1), pp. 635-638,1984
[69] J. Bertrand and P. Bertrand, "Time-frequency representations of broad- band signals" in: Wavelets,
time frequency methods and phase space, edited by J. M. Coombes, A. Grossmann and Ph.
Tchamitchian, Springer-Velag,1989.
[70] J. Bertrand, P. Bertrand, and J. F Ovarlez, "Discrete Mellin transform for signal analysis," Proc.
IEEE ICASSP- 90, pp. 1603-1606,1990.
[71J J. Bertrand and P. Bertrand Affine Time-Frequency Distributions," in: Time-Frequency Signal Anal-
ysis, edited by B. Boashash, pp. 118-140, Longman Cheshire, 1992.
[72] J. Bertrand and P Bertrand, A class of affine Wigner functions with extended covariance proper-
ties," Jour. Math. Phys., vol. 33, pp. 2515-2527,1992.
[73] P Bertrand, J. P. Doremus, B. lzrar, V. T Nguyen and M. R. Feix, "Obtaining non-negative quantum
distribution functions" Phys. Lett., vol. 94, pp. 415-417,1983.
[74] P. Bertrand and J. Bertrand "Time-Frequency Representations of Broad Band Signals," in: The
Physics of Phase Space, edited by Y. S. Kim and W. W. Zachary, Springer Verlag, pp. 250-252,1987.
[75] T K. Bhattacharya, and S. Hayldn, "Application of Time-Frequency Distributions for Improved
Radar Target Detection," Proc. IEEE-SP Inf. Sympos. Time-Frequency and Time-Scale Analysis, pp.
509-512,1992
[76] A. Blanc-Lapierre and B. Picinbono, "Remarques sur la Notion de Spectre instantane de puis-
sance," Public. Scn. Univ. d'Alger, Serie B1, pp. 2-32,1955.
[771 B. Boashash, L. B. White, and J. Imberger, "Wigner-Ville Analysis of Non-Stationary Random
Signals with Application to Temperature Gradient Microstructure," Proc. IEEE ICASSP- 86, pp
2323-2326,1986.
[78] B. Boashash, B. Lovell, and L. B. White, "Time-frequency analysis and pattern recognition using
singular value decomposition of the Wigner-Ville distribution," Proc. SPIE, vol. 826, pp. 104-114,
1987.
[791 B. Boashash, "Note on the use of the Wigner distribution for time-frequency signal analysis," IEEE
Trans. Acoust., Speech, Signal Processing, vol. 36, pp. 1518-1521,1988.
[80] B. Boashash, G. Jones and P. O'Shea, "Instantaneous frequency of signals, estimation techniques
and applications," Proc. SPIE, vo1.1152, pp. 382-4M, 1989.
181] B. Boashash and P O'Shea, A Methodology for Detection and Classification of Some Underwater
Acoustic Signals Using Time-Frequency Analysis Techniques," IEEE Trans. Acoust., Speech, Signal
Processing, vol. 38, pp 1829-1841,1990
[82] B. Boashash and L. B. White, "Instantaneous frequency estimation and automatic time-varying
filtering," Proc. IEEE ICASSP- 90, pp. 1221-1224,1990.
(83] B. Boashash, "Time-frequency signal analysis," in: Advances in Spectrum Estmuition and Array Pro-
cessing, Ed. S. Haykin, Prentice Hall, pp. 418-517,1990.
[84] B. Boashash and A. P. Reilly, "Algorithms for time-frequency signal analysis," in: Time-Frequency
Signal Analysis, edited by B. Boashash, pp.163-181, Longman Cheshire, 1992.
[85] B. Boashash and G. Jones, "Instantaneous frequency and time-frequency distributions," in: Time-
Frequency Signal Analysis, edited by B. Boashash, pp. 43-73, Longman Cheshire, 1992.
[86] B. Boashash, "Estimating and interpreting the instantaneous frequency of a signal - Part 1: Fun- .
damentals," Proceedings of the IEEE, vol. 80, pp. 520-538,1992.
[87] D. Bohm, Quantum Mechanics, Prentice-Hall, 1951.
[88] P. Boles and B. Boashash, "The cross Wigner-Ville distribution - A two dimensional analysis
method for the processing of vibrosis seismic signals," Proc. IEEE ICASSP- 88, pp. 904-907,1988.
[89] P. Boles and B. Boashash, "Application of the cross Wigner-Ville distribution to seismic surveying,"
in: Time-Frequency Signal Analysis, edited by B. Boashash, pp. 445-966, Longman Cheshire, 1992
[90] P. Bopp, "La Mechamque Quantique Est-Elle Une Mechanique Statistique Classique Particuliere?,"
Ann. I;lnst. H. Poincare, voi.15, pp. 81-112,1956.
Bibliography 267
[91] M. Born and P Jordan, "Zur Quantenmechanik," Zeit. f Phys., vol. 34, pp. 858-888,1925.
[92] G. F. Boudreaux-Bartels and T. W. Parks, "Reducing Aliasing in the Wigner Distribution Using
Implicit Spline Approximation," Proc. IEEE ICASSP- 83, pp. 1438-1441,1983.
[93] G. F Boudreaux-Bartels and T W. Parks, "Signal estimation using modified Wigner distributions,"
Proc. IEEE ICASSP- 84, pp. 22.3.1-22.3.4.1984.
[94] G. F. Boudreaux-Bartels, "Time-varying signals processing using the Wigner distribution time-
frequency signal representation," in: Advances in Geophysical Data Processing, Edited by M. Simaan,
JAI Press, pp. 33-79,1985.
[95] G. F Boudreaux-Bartels and I W Parks, "Time-Varying Filtering and Signal Estimation Using
Wigner Distribution Synthesis Techniques," IEEE Trans. on Acoust, Speech, Signal Processing, vol.
34, pp. 442-X151,1986.
[96] G. F. Boudreaux-Bartels, "Wigner distribution analysis of acoustic well logs,' Proc. IEEE ICASSP-
87, vol. 4, pp. 2237-2240,1987.
[97] G. F. Boudreaux-Bartels and A. Papandreou, "A Generalization of the Choi-Williams Exponential
Distribution," 1991 Asilomar Conference on Signals and Systems, pp. 364-369,1991.
[98] A. C. Bovik, P Maragos, and T F Quatieri, "Measuring Amplitude and Frequency Modulations in
Noise Using Multiband Energy Operators," Proc IEEE-SP Int. Sympos. Time-Frequency and Time-
Scale Analysis, pp. 4-6, 1992.
[99] A. C. Bovik, F Maragos, and I F. Quatien, AM-FM Energy Detection and Separation in Noise
Using Multiband Energy Operators," IEEE Trans Signal Processing, vol. 41, pp 3245-3265,1993.
[100] H. Brand, R. Graham and A. Schenzle, "Wigner distribution of bistable steady states in optical
subharmonic generation," Opt Commun., vol. 32, pp. 359-363,1980.
[101] B. R. Breed and T E. Posch, "A range and azimuth estimator based on forming the spatial Wigner
distribution," Proc. IEEE ICASSP- 84, vol. 3, pp. 418.9.1-2,1984.
[102] K H. Brenner and A. W. Lohmann, "Wigner distribution function display of complex 1D signals,"
Opt. Commun., vol. 42, pp. 310-314,1982.
[103] T W Brotherton, I Pollard, R. Barton, A. Krieger, and S. L. Marple, Application of Tune-Frequency
and Time-Scale Analysis to Underwater Acoustic Transients," Proc. IEEE-SP Int. Sympos. Time-
Frequency and Time-Scale Analysis, pp. 513-516,1992.
[104] T Brotherton, T Pollard, and D. L. Jones, "Applications of time-frequency and time-scale rep-
resentations to fault detection and classification," Proc. IEEE-SP Int. Sympos. Time-Frequency and
Time-Scale Analysis, pp. 95-98,1992.
[105] N. Bukhrain, V. Y. Petrunkin, S. A. Rogov, S. V. Rozov, V. G. Samsonov, and A. G. Tkachenko,
Acoustoelectronic system for generating a modified Wigner distribution," Optoelectronics, Instru-
mentatum, and Data Processing, vol. 6, pp. 51-54,1987.
[106] R. J. Bulgnn, B. J. Rubal, C. R. Thompson, and J. M. Moody, "Comparison of Short-Tune Fourier,
Wavelet and Tune-Domain Analysis of Intracardiac Sounds," Procedings of the 30th Ann. Rocky
Mountain Bionengineering Symposium, vol. 29, pp. 465-472, 1993.
[1071 J. R. Carson and T C. Fry, "Variable frequency electric circuit theory with application to the theory
of frequency-modulation," Bell System Tech. J., vol. 16, pp. 513-540,1937.
[108] N. Cartwright, "Correlations without joint distributions in quantum mechanics," Found. of Phys.,
vol. 4, pp. 127-136, 1974.
[109] N. Cartwright, "A non-negative quantum mechanical distribution function," Physica, vol. 83A, pp.
210-212,1976.
[110] L. F. Chaparro, A. El-Jaroudi, and S. A. Kayhan, "Evolutionary Spectral Estimation on Non-
Stationary Signals", in: Fundamentals of Discrete-Time Systems, Editors: M. Jamshidi, M. Mansour,
B. Anderson, and N. Bose, TSI Press, pp. 219-226,1993.
[111] H. Chen and G. F Boudreaux-Bartels, "On the Application of Two-Dimensional AR Modeling
of the Wigner Distribution for Efficient Coding, Filtering, and Pattern Recognition," Proc. IEEE
ICASSP- 90, pp. 2463-2466,1990.
[112] S-C. Chen, and X. Yang, "Speech recognition with high recognition rate by smoothed spaced
pseudo Wigner-Ville distribution (SSPWD) and overlap slide window spectrum methods," Proc.
IEEE ICASSP- 88, vol. 1, pp. 191-194,1988.
268 Bibliography
[113] D. B. Chester, F. J. Taylor and M. Doyle, "Application of the Wigner distribution to speech process-
ing," IEEE Acoust., Speech and Signal Processing Spectrum Estimation Workshop II, pp. 98-102,1983
[114] D. B. Chester, F J. Taylor and M. Doyle, "On the Wigner distribution," Proc. IEEE ICASSP- 83, vol.
2, pp. 491-494,1983.
[115] D. B. Chester, F J. Taylor and M. Doyle, "The Wigner distribution in speech processing applica-
tions," Jour. Franklin Inst., vol. 318, pp. 415-430, 1984.
[116] P R. Chitrapu and D. J. Waldo, "Tune-Frequency Analysis and Linear Prediction of Cardiac Late
Potentials," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 227-230,1992.
[117] H. I. Choi and W J. Williams, "Improved time-frequency representation of multicomponent
signals using exponential kernels," IEEE Trans on Acoust., Speech, Signal Processing, vol. 37, pp.
862-871,1989.
[118] T A. C. M. Claasen and W. F. G. Mecklenbrauker, "The Wigner distribution - A tool for time-
frequency signal analysis - Part 1: continuous-time signals," Philips Jour, of Research, vol. 35, pp.
217-250,1980.
[119] T. A. C. M. Claasen and W. F. G. Mecklenbrauker, "The Wigner distribution - a tool for tme-
frequency signal analysis - Part 11: Discrete time signals," Philips Jour. Research., vol. 35, pp. 276-300,
1980.
[120] T A. C. M. Claasen and W. F. G. Meddenbrauker, The Wigner distribution - a tool for time-
frequency signal analysis - Part III: Relations with other time-frequency signal transformations,"
Philips Jour. Research., vol. 35, pp. 372-389,1980
[121] T A. C. M. Claasen and W. F G. Mecldenbrauker, "Tune-frequency signal analysis by means of the
Wigner distribution," Proc. IEEE ICASSP- 81, vol. 1, pp. 69-72,1981.
[122] T A. C. M. Claasen and W. F. G. Mecklenbri uker, "The aliasing problem in discrete time Wigner
distributions," IEEE Trans. Acoust., Speech, Signal Processing, vol. 31, pp. 1067-1072,1983.
[123] T A. C M. Claasen and W. F G. Mecklenbrauker, "On the time-frequency discrimination of en-
ergy distributions: can they look sharper than Heisenberg?," Proc. IEEE ICASSP- 84, pp. 41B.7.1-
41B.7.4,1984.
[124] F S. Cohen, G. F. Boudreaux-Bartels and S. Kadambe, "Tracking of unknown non stationary chirp
signals using unsupervised clustering in the Wigner distribution space," Proc. IEEE ICASSP- 88,
pp. 2180-2183,1988.
[125] L. Cohen, "Generalized phase-space distribution functions," Jour. Math. Phys., vol. 7, pp. 781-786,
1966.
[126] L. Cohen, "Quantization Problem and Variational Principle in the Phase Space Formulation of
Quantum Mechanics," Jour. Math. Phys, vol. 17, pp. 1863-1866,1976.
[127] L. Cohen and Y. Zaparovanny, "Positive Quantum Joint Distributions," J. Math. Phys., vol. 21, pp.
794-796,1980.
[128] L. Cohen, and T Posch, "Generalized Ambiguity Functions," Proc. IEEE ICASSP-85, pp. 1033-1036,
1985.
[129] L. Cohen and T Pbsch, "Positive time-frequency distribution functions," IEEE Trans. ASSP, vol. 33,
pp. 31-38,1985.
[130] L. Cohen, "Positive and Negative Joint Distributions," in: Non-Equil br:um Quantum Statistical
Physics, eds. G. Moore and M. O. Scully, Plenum Publishing, 1986.
[131] L. Cohen, "Joint Quantum Probabilities and The Uncertainty Principle," in. New Techniques and
Ideas in Quantum Measurement theory, edited by D. Greenberger, New York Academy of Sciences,
1986
[132] L. Cohen, "A critical review of the fundamental ideas of joint time-frequency distributions," IEEE
Proc. Int. Symp. on Circuits and Systems, vol. 1, pp 28-33,1986.
[133] L. Cohen, "Wigner distribution for finite duration or band limited signals and limiting cases," IEEE
Than. Acoust., Speech, Signal Processing, vol. 35, pp. 796--806,1987.
[134] L. Cohen, "On a fundamental property of the Wigner distribution," IEEE Trans. Acoust., Speech,
Signal Processing, vol. 35, pp. 559-561,1987.
[135] L. Cohen and C. Lee, "Instantaneous Frequency, Its Standard Deviation and Multicomponent
Signals," Proc. SPIE, vol. 975, pp. 186-208,1988
Bibliography 269
[1361 L. Cohen and C. Lee, "Instantaneous Mean Quantities in Time-Frequency Analysis," Proc IEEE
ICASSP- 88, 2188-2191,1988.
[137] L. Cohen, "Time-Frequency Distributions - A Review," Proc. of the IEEE, vol. 77, pp. 941-981,1989.
[1381 L. Cohen and C. Lee, "Standard deviation of instantaneous frequency," Proc. IEEE ICASSP- 89,
pp. 2238-2241,1989.
[139] L. Cohen and C. Lee, "Local bandwidth and optimal windows for the short time Founer trans-
form," Proc. SPIE, vol. 1152, pp. 401-425,1989.
[140] L. Cohen and C. Lee, "Optimal Windows For The Short Time Founer Transform," ISSPA 90, vol. 2,
pp. 912-915,1990.
[141] L. Cohen, "Distributions Concentrated Along The Instantaneous Frequency," Proc. SPIE, vol. 1348,
pp. 149-157,1990.
[1421 L Cohen and C. Lee, "Instantaneous bandwidth for signals and spectrogram," Proc. IEEE ICASSP-
90, pp. 2451-2454,1990.
[143] L. Cohen, "A General Approach For Obtaining Joint Representations in Signal Analysis And An
Application to Scale," Proc. SPIE, vol. 1566, pp. 109-133,1991.
[144] L. Cohen, "A Primer on Time Frequency Distributions," in: Time-Frequency Signal Analysis, edited
by B. Boashash, pp. 3-42, Longman Cheshire, 1992.
[145] L. Cohen and C. Lee, "Instantaneous bandwidth," in: Time-Frequency Signal Analysis, edited by B.
Boashash, pp. 98-117, Longman Cheshire, 1992.
[146] L. Cohen, "Scale and Inverse Frequency Representations," in: The Role of Wavelets in Signal Pro-
cessing Applications, edited by: B W. Suter, M. E Oxley, and G. T Warhola, AFIT/EN-TR-92-3, pp.
118-129,1992.
[147] L. Cohen, "What is a Multicomponent Signal?," Proc. IEEE ICASSP- 92, vol. 5, pp. 113-116,1992.
[148] L. Cohen, "A General Approach For Obtaining Joint Representations in Signal Analysis," Proceed-
ings of the IEEE, submitted, May 1992.
[149] L. Cohen, K Assaleh, and A. Fineberg, "Instantaneous Bandwidth and Formant Bandwidth" in:
Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 13-17,1992.
[1501 L. Cohen, "Instantaneous Anything," Proc. IEEE ICASSP- 93, vol. 5, pp. 105-108,1993.
[151] L. Cohen, "The scale representation," IEEE Trans. on Signal Processing, vol. 41, pp. 3275-3293,1993.
[152] N. M. Cole and D. T Barry "Muscle sounds frequencies ofthe frog are modulated by skeletal
muscle tension," Biophys. J., vol. 66, pp. 1104-114,1994.
[153] T D. Cole and A. S. El-Dinary, "Estimation of Target Vibration form Laser Radar Backscatter Using
Time-Frequency Distributions," in: Proc. SPIE, vol. 1936,1993.
[1541 M. Conner and L. Yao, "Optical generation of the Wigner distribution of 2-D real signals," Appl.
Opt., vol. 24, pp. 3825-3829,1985.
[155] C. E. Cook and M. Bernfield, Radar Signals, Academic Press, 1967.
[156] A H. Costa and G. F. Boudreaux-Bartels, "Optimal Tine-Frequency Representations Using a Mul-
tiform, Tiltable Exponential Kernel," IEEE Trans. Signal Processing, submitted, 1993
[157] G. Cristobal, J. Bescos and J. Santamaria, "Application of the Wigner distribution for image repre-
sentation and analysis," Proc. IEEE Int. Assoc Pattern Recognition; Eighth International Conference on
Pattern Recognition, pp. 998-1000,1986.
[158] G. Cristobal, J Bescos, J. Santamaria and J. Montes, "Wigner distribution representation of digital
images," Patt. Rec. Lett., vol. 5, pp. 215-221,1987.
[159] G. Cristobal, J. Bescos and J. Santamaria, "Local image filtering and image classification through
the Wigner distribution," Proc. SPIE, vol. 813, pp. 377-378,1987.
[1601 G. Cristobal, J. Bescos and J. Santamaria, "Image analysis through the Wigner distribution,' Appi
Opt., vol. 28, pp. 262-271, 1989.
[161] G. Cristobal, C. Gonzalo and J. Bescos, "Image filtering and analysis through the Wigner distribu-
tion," in: Advances in Electronics and Electron Physics, vol. 80, pp. 309-397, Academic Press, 1991.
[162] G. S. Cunningham and W. J. Williams. "Fast Implementation of Time-Frequency Distributions,"
Proc. IEEE-SP Int Sympos. Time-Frequency and Time-Scale Analysis, pp. 241-244,1992.
270 Bibliography
[163] G. S. Cunningham and W. J Williams. "Fast Computation of the Wigner Distribution for Finite-
Length Signals," " Proc. IEEE ICASSP- 92, pp. 23-26,1992.
[164] G. S. Cunningham and W J Williams. "Decomposition of Time-Frequency Distribution Kernels,"
Proc. SPIE, vol. 1770, pp. 351-357,1992.
[165] G. S Cunningham and W. J. Williams "High-Resolution Signal Synthesis for Time-Frequency
Distributions," Proc. IEEE ICASSP- 93, vol. 4, pp. 400-403,1993.
[166] G. S. Cunningham and W. J Williams, "Kernel Decomposition of Time-Frequency Distributions,"
IEEE Trans. on Signal Processing, vol. 42, pp. 1425-1442,1994.
[167] G. S. Cunningham and W. J. Williams, "Fast Implementations of Generalized Discrete Tune-
Frequency Distributions," IEEE Trans. on Signal Processing, vol. 42, p.1496-1508,1994.
[168] I. Daubechies, "Time - Frequency Localization Operators: A Geometric Phase Space Approach,"
IEEE Trans. Inform. Theory, vol. 34,1988, pp. 605-612,1988.
[169] I. Daubechies, "The wavelet transform, time-frequency localization and signal analysis," IEEE
Trans on Information Theory, vol. 36, pp. %1-1005,1990.
[170] I. Daubechies, Ten Lectures on Wavelets, SIAM, 1992.
[171] P Davis and M.G. Amin, "Frequency tracking: adaptive filtering vs. time-frequency distributions,"
Proc. IEEE ICASSP- 91, pp. 3225-3228,1991.
[172] C. Detka, A. El Jaroudi and L. F. Chaparro, "Relating the Bilinear Distributions and the Evolution-
ary Spectrum," Proc. IEEE ICASSP- 93, vol. 4, pp. 496-499,1993.
[173] C. S Detka and A. El-Jaroudi, "The Transitory Evolutionary Spectrum," in: Proc. IEEE ICASSP- 94,
1994.
[174] C. Detka, P. Loughlin and A. El-Jaroudi, "On combining evolutionary spectral estimates," Seventh
SP Workshop on Statistical Signal and Array Processing, pp. 26-29,1994
[175] E. J. Diethorn, "The Generalized Exponential Tune-Frequency Distribution," AT&T Bell Laborato-
rtes Tech. Memo., 1992.
[176] E. J. Diethom, "The Generalized Exponential Time-Frequency Distribution," IEEE Trans. on Signal
Processing, vol. 42, p. 1028-1037,1994.
[177] J. H. Eberly and K Wodkiewicz, "The time-dependent physical spectrum of light," J. Opt. Soc. Am.,
vol. 67, pp 1152-1261,1977.
[178] G. Eichmann and B. Z. Dong, "Two-dimensional optical filtering of 1-D signals," Appl. Opt., vol.
21, pp. 3152-3156,1982.
[179] G. Eichmann and N. M. Marinovich, "Scale-invariant Wigner distribution and ambiguity func-
tions," Proc. SPIE, vol. 519, pp. 18-24,1985.
[180] A. S. El-Dinary and T D. Cole, "Non-Recursive FM Demodulation of Laser Radar Backscatter
Using Time-Frequency Distributions," Proc. IEEE ICASSP- 94, vol. 4, pp. 353-356,1994.
[181] J. Fang, L. E. Atlas, and G. Bernard, "Advantages of Cascaded Quadratic Detectors for Analysis
of Manufacturing Sensor Data," Proc. IEEE-SP Int. Sympos. Time-Frequency and rime-Scale Analysis,
pp. 345-348,1992.
[182] R. M. Fano, "Short-time autocorrelation functions and power spectra," J. Acoust. Soc. Am., vol. 22,
pp. 546-550,1950.
[183] P D. Finch and R. Groblicki, "Bivariate probability densities with given margins," Found. of Phys.,
vol. 14, pp. 549-552,1984.
[184] L. M. Fink, "Relations Between The Spectrum And Instantaneous Frequency Of A Signal," Trans-
lated in: Problems Inform Transm., vol. 2, pp. 11-21,1966.
[185] P Flandrin, "Some features of time-frequency representations of multicomponent signals," Proc.
IEEE ICASSP- 84, pp. 41. B. 4.1-41. B. 4.4,1984.
[186] P Flandrin, Sageloli, J., Sessarego, J. P., and Zakharia, M., "Application of time-frequency analysis
to the characterization of surface waves on elastic targets," Acoustics Letters, vol. 10, pp. 23-28,1986
[187] P Flandrin, "When is the Wigner-Ville spectrum non-negative?," Signal Processing III: Theories and
Applications, vol. 1, pp. 239-242, Elsevier, 1986.
[188] P Flandrin, "On detection-estimation procedures in the time-frequency plane," Proc. IEEE
ICASSP- 86, vol. 4, pp. 2331-2334,1986.
Bibliography 271
[189] P Flandrin, "On the positivity of the Wigner-Ville spectrum," Signal Processing, vol. 11, pp. 187-189,
1987.
[190] P Flandrin, "Tune-frequency receivers for locally optimum detection, Proc. IEEE ICASSP- 88, vol.
5, pp. 2725-2728,1988.
[191] P Flandrin, "A time-frequency formulation of optimum detection," IEEE Trans. on Acoust., Speech,
Signal Processing, vol. 36, pp. 1377-1384,1988.
[192] P. Flandrin, "Maximum signal energy concentration in a time-frequency domain," Proc. IEEE
ICASSP- 88, vol. 4, pp. 2176-2179,1988.
[193] P Flandrin, "Time-frequency and time-scale," Proceedings of the Fourth Annual ASSP Workshop on
Spectrum Estimation and Modeling, pp. 77-80,1988.
[194] P. Flandrin, "Non-destructive evaluation in the time-frequency domain by means of the Wigner-
Ville distribution," Signal Processing and Pattern Recognition in Nondestructive Evaluation of Materials,
ed. C. H. Chen, Springer-Verlag, pp. 109-116,1988.
[195] P Flandrin, "Some aspects of non-stationary signal processing with emphasis on time-frequency
and time-scale methods," Wavelets: Time-Frequency Methods and Phase Space, J. M. Combes, A.
Grossman, and P Tchamitchian, eds., pp. 68-98, Springer Verlag, 1989.
[1%] P Flandrin, "Scale-Invariant Wigner Spectra and Self-Similarity" in: Signal Processing V - Theories
and Applications, L. Torres et al. eds., pp. 149-152, Elsevier,1990.
[197] P Flandrin and O. Rioul, "Affine smoothing of the Wigner-Ville distribution," Proc. IEEE ICASSP-
90, pp. 2455-2458,1990.
[198] P. Flandrin, "On the Time-Scale Analysis of Self Similar Processes," The Role of Wavelets in Signal
Processing Applications, edited by: B. W. Suter, M. E. Oxley, and G. T Warhola, AFIT/EN-TR-92-3,
pp. 137-150,1992.
[199] J. R. Fonollosa and C. L. Nikias, "Wigner Pblyspectra: Higher-Order Spectra in Time Varying Sig-
nal Processing," Proc. IEEE ICASSP- 91, pp. 3085-3088,1991.
[200] J. R. Fonollosa and C. L. Nikias, "Wigner Higher Order Moment Spectra: A review," Proc. SPIE,
vol.1770, pp. 280-289,1992.
[201] J. R. Fonollosa and C. L. Nikias, "Wigner Higher-Order Moment Spectra: Definition, Properties,
Computation and Application to Transient Signal Analysis," IEEE Trans. Signal Processing, vol. 41,
pp. 245-266,1993.
[202] J. R. Fonollosa and C. L. Nikias, "A new positive time-frequency distribution," Proc. IEEE ICASSP-
94, vol. 4, pp. 301-304,1994.
[203] B. D. Forrester, "Use of the Wigner-Vine distribution in helicopter fault detection," Australian Symp.
on Signal Processing and Its Applications - 89, pp. 78-82,1989.
[204] B. D. Forrester, "Time-frequency analysis in machine fault detection," in: Time-Frequency Signal
Analysis, edited by B. Boashash, pp. 406-423, Longman Cheshire, 1992.
[205] M. L. Fowler, L. H. Sibul, "A Unified Formulation for Detection Using Time-Frequency and Time-
Scale Methods," Asilomar Conf on Sig., Syst. and Comp., pp. 637-642,1991.
[206] W L. J. Fox, J. C. Luby, J. W. Pitton, P J. Laughlin, and L. E. Atlas, "Sonar and Radar Range Doppler
Processing Using Cone-Shaped Kernel Time-Frequency Representation," 24th Asilomar Conference
on Signals, Systems, and Computers, pp. 1079-1083,1990.
[207] B. Friedlander and B. Pbrat, "Detection of transient signals by the Gabor representation," IEEE
Trans. Acoust., Speech, Signal Processing, vol. 37, pp. 169-180,1989.
[208] D. H. Friedman, "Instantaneous-frequency distribution vs. time: an interpretation of the phase
structure of speech," Proc. IEEE ICASSP- 85, vol. 3, pp. 1121-1124,1985.
[209] C. Frishberg, "Wigner distribution formulation of the crystallography problem," in: Time-
Frequency Signal Analysis, edited by B. Boashash, pp. 488-497, Longman Cheshire, 1992.
[210] D. Gabor, "Theory of communication," Journal of the LEE, vol. 93, pp. 429-457,1946.
[211] W A Gardner, Statistical Spectral Analysis - A Non-Probabilistic Theory, Prentice-Hall, Englewood
Cliffs, 1988.
[212] H. Garudadri, M. P. Beddoes, J. H. V. Gilbert and A. P Benguerel, "Identification of invariant acous-
tic cues in stop consonants using the Wigner distribution," Proc. Int. Symp. on Applied Signal Pro-
cessing and Digital Filtering LASTED 85, ed. M. H. Hamza, Acta Press, pp. 196--200,1985.
272 Bibliography
[213] H. Garudadn, M. P. Beddoes, A. P. Benguerel and J. H. V. Gilbert, "On computing the smoothed
Wigner distribution," Proc. IEEE ICASSP- 87, pp. 1521-1524,1987
[214] C. G. Gaunaurd, and H. C. Strifors "Frequency and time domain analysis of transient resonance
scattering resulting from the interaction of a sound pulse with submerged elastic shells," IEEE
Trans. on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 40, pp. 313-324,1993.
[215] C G. Gaunaurd, and H. C. Strifors, "Signal processing of ideal acoustic echoes from submerged
shells in the time-frequency domain, via Wigner-type distributions," Proc. Intern. Conference on
Theoretical and Computational Acoustics, D. Lee et. all., editors, World Scientific, 1994.
[2161 N. L. Gerr, "Introducing a Third-Order Wigner Distribution," Proc. of the IEEE, vol. 76, pp. 290-292,
1988.
[217] G. B. Giannakis and A. V. Dandawate, "Consistent Kth-Order Time-Frequency Representations
for (Almost) Cyclostationary Signals," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale
Analysis, pp. 123-126,1992.
[2181 C Gonzalo, J. Bescos, L. R. Berriel-Valdos and J. Santamaria, "Space-variant digital filtering
through the Wigner distribution," Signal Processing IV Theories andApplicatwns, Eds. J. L. Lacoume,
A. Chehikian, N. Martin and J. Malbos, Elsevier, vol. 2, pp. 639-642,1988.
[219] C. Gonzalo, J. Bescos, L. R. Berrel-Valdos and J. Santamana, "Space-variant filtering through the
Wigner distribution function," " Applied Optics, vol 28, pp. 730-736,1989
[220] C. Gonzalo and J. Bescos, "Digital implementation of the 4D Wigner distribution function: Ap-
plication to space variant processing of real images," Signal Processing V. Theories and Applications,
Eds. L. Torres, E. Masgrau and M. A. Lagunas, Elsevier, vol. 1, pp. 717-720,1990.
[221] O. D. Grace, "Instantaneous power spectra," J. Acoust. Soc Am., vol. 69, pp. 191-198,1981.
[222] J E. Gray, T R. Rice, J. E. Conte, and G. Watson, "A Method for Determining the Non-linear Time
Dependence of Windowed Data Gathered from Scattered Signals," Proc. IEEE-SP Int. Sympos
Time-Frequency and Time-Scale Analysis, pp. 525-528,1992.
[273] Y. Grenier, "Parametric Time-Frequency Representations," in: Signal Processing, Eds. J. L. Lacoume,
T S. Durrani and R. Stora, Elsevier, pp. 339-397,1987.
[224] C. Griffin, P Rao, and F. J. Taylor, "Roundoff error analysis of the discrete Wigner distribution using
fixed-point arithmetic," Proc. IEEE ICASSP- 89,
[225] C. Griffin and A. H. Nuttall, "Comparison of two Kernels for the Modified Wigner Distribution
Function," Proc. SPIE, vol. 1566, pp. 439-451,1991.
[226] D. W. Griffin and J. S. Lim, "Signal estimation from modified short-time Fourier transform," Proc
IEEE ICASSP- 83, pp. 804-807,1983.
[227] M. F. Griffin, A. M. Finn, and S. Tulpule, "Diagnostic Signal Analysis Using Rank Reduction and
The Wigner Distribution," in: Proc. 24th Asilomar Conference on Signals, Systems, and Computers, vol.
2,1068-107,1990
[228] M. F. Griffin and S. Tulpule, "Time-Frequency Distributions for Propulsion-System Diagnostics,"
Proc. SPIE, vol. 1566, pp. 179-189,1991.
[229] M. F. Griffin, A. M. Finn, and S. Tulpule, "Signal Enhancement for Diagnostic Signal Analysis'
Proc IEEE ICASSP- 91,, 3281-3284,1991
[230] L. Griffiths, "Rapid measurement of digital instantaneous frequency," IEEE Trans. Acoust., Speech,
Signal Processing, vol. 23, pp. 207-222,1975.
[231] R. Groblicki, "A generalization of Cohen's theorem," ]our. Math. Phys., vol. 24, pp. 841-843,1983.
[232] D. Groutage, J. Schemp and L. Cohen, "Characterization and Analysis of Marine Sounds Using
Time-Frequency and Time-Prony Techniques," in: OCEANS 94 OSATES, to appear, 1994
[233] D. Groutage and D. Bennink, "Analysis of Resonant Structures Using Advanced Time-Prony Tech-
niques," ASME Journal of Acoustics and Vibration, submitted, 1994
[234] Z. Guo, L.-G. Durand and H. C. Lee, "Comparison of Time-Frequency Distribution Techniques for
Analysis of Simulated Doppler Ultrasound Signals of the Femoral Artery" IEEE Trans. on Biomed
Engr., vol. 41, 332-342, 1994.
[235] Z. Guo, L.-G. Durand and H. C. Lee, "The Time-Frequency Distributions of Nonstationary Signals
Based on a Bessel Kernel," IEEE Trans. Signal Processing, vol. 42, pp. 1700-1707,1994
Bibliography 273
[2361 M. S. Gupta, "Definitions of instantaneous frequency and frequency measurability," Amer. Jour
Phys., vol. 43, pp. 1087-1088,1975.
[237] A. K Gupta, T Iwai, and T Asakura, "Parallel generation of the sectional Wigner distribution func-
tion for two-dimensional objects," Jour of Optics, vol. 16, pp. 10-19,1987.
[2381 J. K Hammond, "Evolutionary spectra in random vibrations," Jour. Roy. Statist. Soc., vol. B 35, pp.
167-188,1983.
[239] J. K Hammond, Y. H. Lao and R. F. Harrison, "Evolutionary spectral density models for random
processes having a frequency modulated structure," Proc. IEEE ICASSP- 83, pp. 261-264, 1983.
[240] J. K Hammond, J. S. Lee and R. F. Harrison, "The relationship between Wigner-V lle and evo-
lutionary spectra for frequency modulated random processes," Proc. IEEE ICASSP- 86, pp. 2327-
2330,1986.
[241] F J. Harris and A. Abu Salem, "Performance comparison of Wigner-Ville based techniques to stan-
dard FM-discriminators for estimating instantaneous frequency of rapidly slewing FM sinusoid
in the presence of noise," Proc. SPIE, vol. 975, pp. 232-244,1989.
[242] F. J. Harris, "Techniques and limitations of spectrum analysis with the discrete Fourier transform,"
in: Time-Frequency Signal Analysis, edited by B. Boashash, pp. 184-207, Longman Cheshire, 1992.
[2431 J. D. Harris and D. Nelson, "Glottal Pulse Alignment in Voiced Speech for Pitch Determination,"
Proc. IEEE ICASSP- 93, pp. 519-522,1993.
[244] S. Hearon and M. G. Amin, "Statistical trade offs in modern t-f distributions," Proceedings of Asilo-
mar Conference on Circuits, Systems, and Computers, vol. 2, pp. 359-363,1991.
[2451 S. Hearon and M. Amin, "Minimum variance time-frequency distribution kernels," IEEE Transac-
tums on Signal Processing, to appear, 1994.
[246] C. W. Helstrom, "An expansion of a signal in Gaussian elementary signals," IEEE Trans. on Informa-
tion Theory, vol. 12, pp. 81-82,1966.
[247] M. Hillery, R. F. O'Connell, M. O. Scully and E. P Wigner, "Distribution functions in physics: Fun-
damentals," Physics Reports, vol. 106, pp. 121-167,1984.
[248] R. Hippenstiel and P. De Oliveira, "Time-varying spectral estimation using the instantaneous
power spectrum (IPS)," IEEE Trans. Sig. Proc., vol. 38, pp. 1752-1759, 1990.
[249] F. Hlawatsch, "Interference terms in the Wigner distribution," in: Digital Signal Processing-84, Eds.
V. Cappellini and A. G. Constantinides, Elsevier, 1984.
[250] F. Hlawatsch, "Transformation, Inversion and Conversion of Bilinear Signal Representations,"
Proc. IEEE ICASSP- 85, pp. 1029-1032,1985
[251] F. Hlawatsch and W. Krattenthaler, "Signal Synthesis from Unitary Tune-Frequency Signal Rep-
resentations," In Signal Processing III: Theories and Applications, eds. I. T. Young et al., pp. 37-40,
Elsevier, 1986.
[252] F. Hlawatsch, "Unitary Tune-Frequency Signal Representations," In Signal Processing III. Theories
and Applications, eds. I. T Young et al., pp. 33-36, Elsevier, 1986.
[2531 a) F. Hlawatsch, "A note on 'Wigner distribution for finite duration or band-limited signals and
limiting cases'," IEEE Trans. Acoust., Speech, Signal Processing, vol. 36, pp. 927-929,1988. b) Author's
reply, ibid.
[254] F Hlawatsch, W Kozek and W Krattenthaler, "Time-frequency subspaces and their application to
time-varying filtering," Proc. IEEE ICASSP- 90, pp. 1607-1610,1990.
[255] F. Hlawatsch, "Duality and Classification of Bilinear Time-Frequency Signal Representations,"
IEEE Trans. Signal Processing, vol. 39, pp. 1564-1574,1991.
[256] E Hlawatsch, A. Papandreou, and G. F Boudreaux-Bartels, "The Hyperbolic Class of Quadratic
Time-Frequency Representations Part II: Regularity and Unitarity, Subclasses, and the Altes-
Marinovich Distribution," IEEE Trans. Signal Processing, submitted, 1992.
[257] F. F lawatsch and G. F. Boudreaux-Bartels, "Linear and Quadratic Time-Frequency Signal Repre-
sentations," IEEE Signal Processing Magazine, vol. 9, pp. 21-67,1992.
[2581 F. Hlawatsch, "Regularity and Unitarity of Bilinear Time-Frequency Signal Representations," IEEE
Trans. Information Theory, vol. 38, pp. 82-94, 1992.
[259] F. Hlawatsch, "Bilinear Signal Synthesis," IEEE Trans. Signal Processing vol. 40, pp. 352-363,1992.
274 Bibliography
(2601 F. Hlawatsch and W. Kozek, "Time-Frequency Weighting and Displacement Effects in Linear,
Time-Varying Systems," Proc IEEE ISCAS-92, pp. 1455-1458, 1992
[261] F. Hlawatsch, A. Papandreou, and G. F. Boudreaux-Bartels, "The Power Classes of Quadratic
Time-Frequency Representations: A Generalization of the Affine and Hyperbolic Classes," Proc.
IEEE ICASSP- 93, pp. 1265-1270,1993
[262] F. Hlawatsch, "Tune-Frequency Design and Processing of Signals via Smoothed Wigner Distribu-
tions," IEEE Trans. Signal Processing, vol. 41, pp. 278-287, 1993.
[263) F Hlawatsch, A Papandreou, and G. F Boudreaux-Bartels, "Regularity and Urutanty of Affine
and Hyperbolic Tune-Frequency Representations," Proc IEEE ICASSP- 93, vol. 3, pp. 245-248,
1993.
[2641 F. Hlawatsch and R. L. Urbanke, "Bilinear Tune-Frequency Representations of Signals: The Shift-
Scale Invariant Class," IEEE Trans Signal Processing, vol. 42, pp. 357-366, 1994
[265) F. Hlawatsch and W. Kozek, "Second-Order Time-Frequency Synthesis of Nonstationary Ran-
dom Processes," IEEE Trans. Info Theory, to appear, 1994
[266] R. L. Hudson, "When is the Wigner quasi-probability density non-negative7," Rep. Math. Phys.,
vol. 6, pp. 249-252, 1974
[267] D. H. Hughes, "Backscattering of Sound by Spherical Shells in Water," Ph. D. dissertation, Wash-
ington State University, 1992.
[268] D. H Hughes and P L. Marston, "Local temporal variance of Wigner's distribution function as a
spectroscopic observable. Lamb wave resonances of a spherical shell," I Acoust. Sac. Am., vol 94,
p. 499,1993.
[269] D. H. Hughes, C. F. Gaumond and L. R. Dragonette, "Synthesized wave packet basis for monostatic
scattering from a randomly ribbed, finite cylindrical shell," J Acoust. Soc. Am., submitted, 1994
[2701 J. Imberger and B. Boashash, "Application of the Wigner-Ville distribution to temperature gradient
nucrostructure: A new technique to study small-scale variations," Jour of Physical Oceanography,
vol. 16, pp. 1997-2012,1986.
[271] M. Jacob and P Fortier, "An Interpretation of the Spectrogram for Discrete-Time Signals," Proc.
IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 281-284,1992.
[272] L. D. Jacobson and H. Wechsler, "The composite pseudo Wigner distribution (CPWD): A com-
putable and versatile approximation to the Wigner distribution (WD)," Proc IEEE ICASSP- 83,
pp. 254-256,1983.
[273] L. D. Jacobson and H. Wechsler, "Joint spatial/spatial-frequency representations," Signal Process-
ing, vol. 14, pp. 37-68,1988.
[2741 Janse, C. F, and Kaizer, A. J. M., "Time-Frequency Distributions of Loudspeakers. The Application
of the Wigner Distribution," J. Audio Eng. Soc., vol. 31, pp. 198- 223,1983.
[275] A. J. E. M. Janssen, "Positivity of weighted Wigner distributions," SIAM Journal of Mathematical
Analysis, vol. 12, pp. 752-758,1981.
[2761 A. J. E. M. Janssen, "On the locus and spread of pseudo-density functions in the time-frequency
plane," Philips Journal of Research, vol. 37, pp. 79-110,1982
[277] A. J. E M. Janssen, "Positivity properties of phase-plane distribution functions," Jour Math. Phys,
vol. 25, pp. 2240-2252,1984.
[278] A. J. E. M. Janssen and T A. C. M. Claasen, "On positivity of time-frequency distributions," IEEE
Trans. Acoust., Speech, Signal Processing, vol. 33, pp. 1029-1032,1985.
[279] A. J. E. M. Janssen, "Bilinear phase-plane distribution functions and positivity," Jour Math. Phys,
vol. 26, pp. 1986-1994,1985.
[280] A. J. E. M. Janssen, "A note on 'Positive time-frequency distributions," IEEE Trans. Acoust., Speech,
Signal Processing, vol. 35, pp. 701-703, 1987. See also: Author' reply, ibid.
[281] A. J. E. M. Janssen, "Positivity of time-frequency distribution functions," Signal Processing, vol. 14,
pp. 243--252,1988
[2821 J. Jeong and W. J Williams, "Variable Windowed Spectrograms: Connecting Cohen's Class and
the Wavelet Transform," ASSP Workshop on Spectrum Estimation and Modeling, pp. 270-273, 1990
Bibliography 275
[283] J. Jeong, G. S. Cunningham and W. J. Williams. "Instantaneous Frequency and Kernel Require-
ments for Discrete Time-Frequency Distributions," Proc SPIE, vol. 1348, pp. 170-180,1990.
[284] J. Jeong and W. J. Williams, "A new formulation of generalized discrete-time time-frequency
distributions," Proc IEEE ICASSP- 91, pp. 3189-3192,1991.
[285] J. Jeong and W. Williams, "Kernel design for reduced interference distributions," IEEE Trans Sig.
Proc., vol. 40, pp. 402-412,1992.
[2861 J. Jeong and W. J. Williams. "Alias-Free Generalized Discrete-Tune Time-Frequency Distributions,"
IEEE Trans. on Signal Processing, vol. 40, pp. 2757-2765, 1992.
[287) J. Jeong and W. J. Williams. "Time-Varying Filtering and Signal Synthesis," in: Time-Frequency
Signal Analysis, edited by B. Boashash, pp. 389-405, Longman Cheshire, 1992.
[288] J. Jeong, W. J. Williams. "A Unified Fast Recursive Algorithm for Data Shuffling in Various Orders,"
IEEE Trans on Signal Processing, vol. 40, pp. 1091-1095,1992.
[2891 J. Jeong and W. J. Williams. "Mechanism of the Cross-Terms in Spectrograms," IEEE Trans. on Signal
Processing, vol. 40, pp. 2608-2613,1992.
[290] D. L. Jones and T W. Parks, "A resolution comparison of several time-frequency representations,"
IEEE Trans. on Signal Processing, vol. 40, pp. 413-420,1992.
[291] D. L. Jones and R. G. Baraniuk, 'An adaptive optimal kernel time-frequency representation,"
Proc. IEEE ICASSP- 93, vol. 4,109-112,1993.
[292] G. Jones and B. Boashash, "Instantaneous frequency, time delay, instantaneous bandwidth and
the analysis of multicomponent signals," Proc. IEEE ICASSP- 90, pp. 2467-2470,1990.
[293] S. Kadambe, G. F. Boudreaux-Bartels, F Duvaut, "Window Length Selection for Smoothing
Wigner Distributions by Applying an Adaptive Filter," Proc. IEEE ICASSP- 89, pp. 52.6.1-4,1989.
[294] S. Kadambe and G. F. Boudreaux-Bartels, "A Comparison of the Existence of 'Cross Terms' in the
Wigner Distribution and the Squared Magnitude of the Wavelet Transform and the Short Time
Fourier Transform," IEEE Trans on Signal Processing, vol. 40, pp. 2498-2517,1992.
[295] J. F. Kaiser, "On a simple algorithm to calculate the'energy' of a signal;' Proc. IEEE ICASSP- 90,
pp. 381-384, 1990.
[296] J. F. Kaiser, "Some Useful Properties of Teager's Energy Operators," Proc. IEEE ICASSP- 93, vol. 3,
pp. 149-152,1993.
[297] S. Kay and G. F. Boudreaux-Bartels, "On the optimality of the Wigner distribution for detection,"
Proc. IEEE ICASSP- 85, pp. 1017-1020,1985.
[298] S. M. Kay, Modern Spectral Estimation, Prentice-Hall, Englewood Cliffs, 1987.
[299] S. A. Kayhan, L. F. Chaparro, and A. El Jaroudi, "Wold-Cramer Evolutionary Spectral Estimators,"
Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 115-118,1992.
[300] A. S. Kayhan, A. El-Jaroudi, L. F. Chaparro, "The Evolutionary Periodogram for Non-Stationary
Signals," IEEE Trans. on Signal Processing, vol. 42, pp. 1527-1536,1994.
[301] A. S. Kayhan, A. El-Jaroudi, and L. F. Chaparro, "Data-Adaptive Evolutionary Spectral Estima-
tion," IEEE Trans. on Signal Processing, to appear, 1994.
[302] E. Kelly and R. P. Wishner, "Matched-Filter Theory Velocity, Accelerating Targets," IEEE Trans on
Mil. Electr., vol. 9, pp. 56-69, 1965.
[303] E. L. Key, E. M. Fowle and R. D. Haggarty, A method of designing signals of large time-bandwidth
product," IRE Intern. Convent. Rec., vol. 4, pp. 146-154,1961.
[304) A. Khmchin, Bull. Univ. Moscow, 1, 1937.
[305] J. G. Kirkwood, "Quantum statistics of almost classical ensembles," Phys. Rev., vol. 44, pp. 31-37,
1933.
[306] J. R. Klauder, A. C. Price, S. Darlington and W. J. Albershein, "The theory and design of chirp
radars," Bell System Tech. J., vol. 39, pp. 745-808,1960.
[307] J. R. Klauder, "Path Integrals For Affine Variables," in: Functional Integration: Theory and Applications,
J. P Antoine and E. Tuapogui, eds. Plenum Press, New York, 1980.
[308] Kobayashi, F., and Suzuki, H., "Tune-varying signal analysis using squared analytic signals," Proc.
IEEE ICASSP- 87, vol. 3, pp. 1525-1528,1987.
276 Bibliography
[309] 1 E. Koczwara and D. L. Jones, "On mask selection for tune-varying filtering using the Wigner
distribution," Proc. IEEE ICASSP- 90, pp. 2487-2490,1990.
[310] K Kodera, C. de Villedary and R. Gendrin, "A new method for the numerical analysis of non-
stationary signals," Physics of the Earth and Planetary Interiors, vol. 12, pp. 142-150, 1976.
[311] K Kodera, R. Gendrin and C. de Villedary, "Analysis of time-varying signals with small BT values,"
IEEE Trans. Acoust., Speech, Signal Processing, vol 26, pp. 64-76,1978.
[3121 R. Koenig, H. K Dunn and L. Y. Lacy, "The sound spectrograph," J. Acoust. Soc. Am., vol. 18, pp.
1919,1946.
[313] W. Kolodziej and D. H. Park, "Detection and Diagnosis of Parameters Change in Linear Systems
Using Time-Frequency Transformation," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale
Analysis, pp. 369-372,1992
[3141 W. Kozek, "Time-Frequency Signal Processing Based on the Wigner-Weyl Framework," Signal
Processing, vol. 29, pp. 77-92,1992.
[315] W. Kozek, "On the Generalized Weyl correspondence and its Application to Time-Frequency
Analysis of Linear Systems," Proc. IEEE-SP Int Sympos. Time-Frequency and Time-Scale Analysis,
pp. 167-170,1992.
[316] W Krattenthaler and F Hlawatsch, "Improved Signal Synthesis from Pseudo Wigner Distribution,"
IEEE Trans. Signal Processing, vol. 39, pp. 506-509, 1991.
[317] W. Krattenthaler and E. Hlawatsch, "Tune-frequency design and processing of signals via
smoothed Wigner distributions," IEEE Trans on Acous., Speech, Signal Processing, vol. 41, pp 278-
287,1993.
[3181 J. G. Kruger and A. Poffyn, "Quantum mechanics in phase space," Physica, vol. 85, pp. 84-100,1976.
[319] B. V. K V. Kumar and W. C. Carroll, "Performance of Wigner distribution function based detection
methods," Optical Engineering, vol 23, pp. 732-737,1984.
[3201 B. V K V. Kumar, C. P. Neuman and K. J. DeVos, "Discrete Wigner synthesis," Signal Processing,
vol. 11, pp. 277-304,1986.
[3211 B. V. K. V. Kumar and K J. DeVos, "Linear system description using Wigner distribution functions,"
Proc. SPIE, vol. 826, pp. 115-124,1987.
[322] R. Kumaresan, A. G. Sadasiv, C. S. Ramalingam, and J. E Kaiser, "Instantaneous Non-linear Oper-
ators for Tracking Multicomponent Signal Parameters," Proc. IEEE-SP Int. Sympos. Time-Frequency
and Time-Scale Analysis, pp. 404-407,1992.
[323] R. Kumaresan and C. S. Ramalingam, "On separating Voiced-Speech into its Components" in:
Proc. of Twenty-Seventh Annual Asilomar Conference on Signals, Systems and Computers, vol. 2, 1041-
1046,1993.
[3241 V. V. Kuryshkin, "Some problems of quantum mechanics possessing a non-negative phase-space
distribution functions," International Journal of Theoretical Physics, vol. 7, pp. 451-466,1973.
[3251 V. V. Kuryshkin, "Uncertainty principle and the problems of joint coordinate-momentum proba-
bility density in quantum mechanics," in: The uncertainty Principle and the Foundations of Quantum
Mechanics, pp. 61-83, Wiley, 1977.
[3261 S. Kwong W. Gang, and O. Y. Zheng, "Fundamental Frequency Estimation Based on Adaptive
Tune-Averaging Wigner- Vdle Distribution," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-
Scale Analysis, pp. 413-416,1992.
[327] U. K Laine, "Famlet, to be or not to be a wavelet?," Proc. IEEE-SP Int. Sympos. Time-Frequency and
Time-Scale Analysis, pp. 335-338,1992.
[328] D. G. Lampard, "Generalization of the Wiener-Khinchin theorem to non-stationary processes,"
Jour. Applied Physics, vol. 25, pp. 802,1954.
[329] H. Landau, D. Slepian and H. Pollack, "Prolate spheroidal wave functions, Fourier analysis and
uncertainty," Bell System Tech. J., vol. 40, pp. 43-63, 65-84, 1961.
[3301 F. Lari and A. Zakhor, "Automatic Classification of Active Sonar Data Using Time-Frequency Trans-
forms," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 21-24,1992.
[331] R. M. Lerner, "Representation of signals," in- Lectures on Communication System Theory, E. J. Bagh-
dady, ed. , pp. 203-242, McGraw-Hill, 1961
Bibliography 277
[332] M. J. Levin, "Instantaneous Spectra and Ambiguity Functions," IEEE Trans. Inform. Theory, vol. 10,
pp. 95-97,1964.
[333] A. L. Levshin, V. F. Pisarenko and G. A. Pogrebinsky, "On a frequency-time analysis of oscillations,"
Ann. Gcophys., vol. 28, pp. 211-218,1972.
[334] Y. Li, G. Eichmann and M. Conner, "Optical Wigner distribution and ambiguity function for com-
plex signals and images," Optics Communications, vol. 67, pp. 177-179,1988.
[335] Z.-b Lin, "Wideband ambiguity function of broadband signals," I Acoust. Soc. Am., vol. 83, 2108-
2116,1988.
[336] Z. Y. Lin and J. De Z. Chen, "Time - Frequency Representation of the Electrogastrogram - Appli-
cation of the Exponential Distribution," IEEE Trans. on Biomed Engr., vol. 41, pp. 267-275,1994.
[337] S. C. Liu, "Time-varying spectra and linear transformation," Bell Systems Tech. I., vol. 50, pp 2365-
2374,1971.
[338] P Loughlin, J. Pitton, L. E. Atlas, "New properties to alleviate interference in time-frequency rep-
resentations," Proc. IEEE ICASSP- 91, pp. 3205-3208,1991.
[339] P. Loughlin, J. Pitton and L. E. Atlas, An information-theoretic approach to positive time-
frequency distributions," Proc. IEEE ICASSP- 92, vol. 5, pp. 125-128,1992.
[340] P Loughlin, J Pitton and L. E. Atlas, "Proper time-frequency energy distributions and the Heisen-
berg uncertainty principle," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp.
151-154,1992.
[341] P Loughlin, L. E. Atlas and J. Pitton, "Advanced time-frequency representations for speech pro-
cessing," in: Visual Representations of Speech Signals, M. Cooke, S. Beete and M. Crawford (eds. ),
John Wiley & Sons, 1993.
[342] P Loughlin, J. Pitton and L. E. Atlas, "Bilinear time-frequency representations: new insights and
properties," IEEE Trans Sig. Proc., vol. 41, pp. 750-767,1993.
[343] P Loughlin, J. Pitton and L. E. Atlas, "Construction of positive time-frequency distributions," IEEE
Trans. Sig. Proc., to appear, 1994.
[344] P Loughlin, J. Pitton and B. Hannaford, "Approximating time-frequency density functions via
optimal combinations of spectrograms, IEEE Sig. Proc. Letters, submitted, 1994.
[345] P Loughlin, "Comments on scale invariance of time-frequency distributions," Signal Processing
Letters, submitted, 1994.
[346] B. Lovell and R. Williamson, "The statistical performance of some instantaneous frequency esti-
mators," IEEE Trans. Sig. Proc., vol. 40, pp. 1708-1723,1992.
[347] B. Lovell, R. Williamson and B. Boashash, "The relationship between instantaneous frequency and
time-frequency representations," IEEE Trans. Sig. Proc., vol. 41, pp.1458-1461,1993.
[348] Loynes, R. M. "On the concept of spectrum for nonstationary processes" J. Roy. Stat. Soc. Ser. B,
vol. 30, pp. 1-30,1968.
[349] D. B. Malkoff and E. J. Smythe, "An integrated neural network-time-frequency method for detec-
tion and classification," in: Proceedings of SPIE 1990 Technical Symposium on Optical Engineering and
Phonetics in Aerospace Sensing, Orlando, Florida, 1990.
[350] D. B. Malkoff, "Detection and classification by neural networks and time distributions," in: Time-
Frequency Signal Analysis, edited by B. Boashash, pp. 324-348, Longman Cheshire, 1992.
[351] S.C. Mallat and Z. Zhang, "Matching Pursuits with Time-Frequency Dictionaries," IEEE Trans. on
Signal Processing, vol. 41, p. 3397-3415,1993.
[352] S. G. Mallat and Z. Zhang, "Adaptive Time-Frequency Decomposition Using Matching Pursuits,"
Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 7-10,1992.
[353] L. Mandel, "Interpretation of instantaneous frequency," Amer. Jour Phys., vol. 42, pp. 840-846,1974.
[354] S. Mann and S. Haykin, "Adaptive "chirplet" transform: An adaptive generalization of the wavelet
transform," Optical Engineering, vol. 31, pp. 1243-1256,1992.
[355] S. Mann and S. Haykin, "Tune-frequency perspectives: the chirplet transform," Proc. IEEE
ICASSP- 92, vol. 3, pp. 417-420,1992.
[356] S. Mann and S. Haykin, "'Chirplets' and'Warblets': Novel Time-Frequency Methods," Electronics
Letters, vol. 28, pp. 114-118,1992,
278 Bibliography
[357] P. Maragos, J. F. Kaiser, and I E Quaheri, On Amplitude and Frequency Demodulation Using
Energy Operators," IEEE Trans Signal Processing, vol. 41, pp. 153Z-1550,1993.
[358] H. Margenau and R. N. Hill, "Correlation between measurements in quantum theory," " Prog. The-
oret Phys., vol. 26, pp. 722-738,1961.
[359] H. Margenau and L Cohen, "Probabilities in quantum mechanics," m: Quantum Theory and Reality,
edited by M. Bunge, Springer-Verlag,1%7.
[360) N. M. Marinovich and W. A. Smith, "Use of the Wigner distribution to analyze the time-frequency
response of ultrasonic transducers," in: Proc. IEEE 1984 Ultrasonics Symposium, 1984.
[361] N. M. Marinovich and W. A. Smith, "Suppression of noise to extract the intrinsic frequency varia-
tion from an ultrasonic echo," Proc. IEEE 1985 Ultrasonics Symposium, pp. 841-846, 1985.
[362] N. M. Marinovich and G. Eichmann, "Feature extraction and pattern classification in space-spatial
frequency domain," Proc. SPIE, vol. 579, pp. 15-20,1985.
[363] N. M. Mannovich and G. Eichmann, An expansion of the Wigner distribution and its applica-
tions,' Proc. IEEE ICASSP- 85, vol. 3, pp. 1021-1024,1985.
[364] N. M. Marinovich, "The Wigner distribution and the ambiguity function: Generalizations, en-
hancement, compression and some applications," Ph. D. Dissertation, The City University of New
York, NY, 1986.
[365] N. M. Marinovich, and W. A. Smith, "Application of joint time-frequency distributions to ultrasonic
transducers," 1986 IEEE Information Symposium on Circuits and Systems, pp. 50-54,1986.
[366] N. M. Marinovich, "VLSI architecture of a real-time Wigner distribution processor for acoustic
signals," Proc. IEEE ICASSP- 88, vol. 4, pp. 2112-2115,1988.
[367] N. M. Marinovitch, "Detection of transients in stationary colored noise by time frequency distnbu-
tions," in: Time-Frequency Signal Analysis, edited by B. Boashash, pp. 305-323, Longman Cheshire,
1992.
[368] W. D. Mark, "Spectral analysis of the convolution and filtering of non-stationary stochastic pro-
cesses," Jour. Sound Vib., vol. 11, pp. 19-63, 1970.
[369] W. D. Mark, "Power spectrum representation for nonstationary random vibration," in: Random
Vibration - Status and recent developments, edited by D. H. lelishakoff, Elsevier,1986.
[370] W. D. Mark, Effects of intensity modulations on the power spectra of random processes," Jour.
Sound and V7i r., vol. 119, pp. 451-485,1987
[371] R. J. Marks II, J. F. Walkup, and T F. Krile, "Ambiguity function display: an improved coherent
processor," Appl. Opt., 16,746-750,1977.
[372] R. J. Marks II, Introduction to Shannon Sampling and Interpolation Theory", Springer-Verlag,1991.
[373] S. L. Marple, Digital Spectral Analysis, Prentice-Hall, 1987.
[374] L. Marple, I Brotherton, R. Barton, et al., "Travels through the time-frequency zone: advanced
Doppler ultrasound processing techniques," Proc. 27th Asilomar Conf on Sigs., Syst. and Comps., vol.
2, pp. 1469-1473,1993.
[375] P L. Marston, L. Zhang, N. Sun, G. Kaduchak, and D. H. Hughes, "Ray representations of the
backscattering of tone bursts by shells in water: Calculations and related experiments," in: Second
International Congress on Recent Developments in Air and Structure-Borne Sound and Vibration, M. J.
Crocker and P K Raju, editors, vol 3,1992.
[376] W. Martin, "Time-frequency analysis of random signals," Proc IEEE ICASSP- 82, pp. 1325-13281
1982.
[377] W. Martin, "Measuring the degree of non-stationarity by using the Wigner-Ville spectrum," Proc
IEEE ICASSP- 84, pp. 41B.3.1-1,1984.
[378] W. Martin and P Flandrin, "Wigner-Ville spectral analysis of non-stationary processes," IEEE
Trans. Acoust., Speech, Signal Processing, vol. 33, pp. 1461-1470,1985.
[379] W. Martin, and P Flandrin, "Detection of changes in signal structure by using the Wigner-Ville
spectrum," Signal Processing, vol. 8, pp. 215-233,1985.
[380] N. H. McCoy, "On the Function in Quantum Mechanics Which Corresponds to a Given Function
in Classical Mechanics," Proc. Natl Acad. Sa. U.S.A., vol. 18, pp. 674-676,1932.
Bibliography 279
[381] 'E J. McHale and G. F. Boudreaux-Bartels, "An Algorithm for Synthesizing Signals from Partial
Time-Frequency Models Using the Cross Wigner Distribution," IEEE Trans. Signal Processing, vol.
41, pp. 1986-1990,1993.
[382] W. F G. Mecklenbrhuker, "A tutorial on non-parametric bilinear time-frequency signal represen-
tations," in: Signal Processing, Eds. J. L. Lacoume, T S. Durrani and R. Stora, Elsevier, 277-338,
1985.
[383] F Molinaro and F. Castanie, "A comparison of time-frequency methods," Signal Processing V. -
Theories and Applications, Eds. L. Torres, E. Masgrau and M. A. Lagunas, Elsevier, vol. 1, pp. 145-
149,1990.
[3841 P M. Moore and Q. T Tran, "Signal-Dependent Kernel for Reduction of Time-Domain Cross Terms
in Tune-Frequency Distributions," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Anal-
ysis, pp. 245-248, 1992.
[385] J. M. Morris and S. Qian, "A Cross-Term Free Joint Tune-Frequency Representation via Wigner-
Vile Distribution Decomposition," Proc. 29th Annual Allerton Conf Commun., Contr., & Computing,
pp. 553-554,1991.
[386] J. M. Morris and D-S Wu, "Some Results on Joint Time-Frequency Representations via Gabor
Expansions," Proc. 1992 Conf. Inform. Sci. & Sys., vol. 1, pp. 6-10,1992.
[387] J. M. Morris and D-S. Wit, "On Alias-Free Formulations of Discrete-Time Cohen's Class of Distri-
butions," IEEE Trans. on Signal Processing, submitted, 1994.
[388] J. M. Morris and D-S. Wu, "On the Resolution of the Crossterm-Deleted Wigner Representation,"
preprint, 1994.
[389] G. Mourgues, M. R. Feix, J. C. Andrieux and P Bertrand, "Not necessary but sufficient conditions
for the positivity of generalized Wigner functions," Jour Math. Phys., vol. 26, pp. 2554-2555,1985.
[3901 J. E. Moyal, "Quantum mechanics as a statistical theory," Proc. Camb. Phil. Soc., vol. 45, pp. 99-124,
1949.
[391] M. Mugur-Schachter, "A study of Wigner's theorem on joint probabilities," Found. of Phys., vol. 9,
pp. 389-404,1979.
[392] S. K Mullick and V. A. Topkar, A Wigner distribution based receiver," " Signal Processing, vol. 14, pp.
185-1%,1988.
[393] W. Music, W. Fox, P Loughlin, L. E. Atlas and J. Pitton, "Shift-keyed signal identification using
time-frequency processing," in: Proc. 25th Asilomar Conf. on Signals, Systems and Computers, vol. 2,
846-850,1991.
[394) F. J. Narkowich, and R. F. O'Connell, "Necessary and sufficient conditions for a phase-space func-
tion to be a Wigner distribution," Phys. Rev. A, vol. 34, pp. 1-6,1986.
[395] S. N. Nawab, I F Quatieri and J. S. Lim, "Signal reconstruction from short-time Fourier transform
magnitude," Proc. IEEE ICASSP- 82, pp. 1046-1048,1982.
[3%) S. H. Nawab, T F Quatieri and J. S. Lim, "Algorithms for signal reconstruction from short-time
Fourier transform magnitude," Proc. IEEE ICASSP- 83, pp. 800-803,1983.
[397] S. N. Nawab and T F Quatieri"Short-Tune Fourier Transform," Advanced Topics in Signal Processing,
ed., J. S. Lim and A. V. Oppenheim, pp. 289-337, Prentice-Hall, 1988.
[3981 S. N. Nawab, D. Beyerbach and E. Dorken, "Principal decomposition of time-frequency distribu-
tion," IEEE Trans. Sig. Proc., vol. 41, pp. 3182-3186,1993.
[399] D. Nelson, D. Barbour, and M. Pierce, "Speech Analysis by Alternate Spectral Analysis Methods,"
preprint, 1994.
[400] D. Nelson, "The Wavelet Mellin Transform," preprint, 1994.
[4011 C. L. Nikias and A. P Petropulu, Higher-Order Spectral Analysis: a Nonlinear Signal Processing Frame-
work, Prentice-Hall, 1993.
[4021 A. H. Nuttall, "On the quadrature approximation to the Hilbert transform of modulated signals,"
Proc. IEEE, vol 54, pp. 1458,1966; and Reply by Bedrosian.
[403] A. H. Nuttall, "Wigner distribution function: Relation to short-term spectral estimation, smooth-
mg, and performance in noise," Naval Underwater Systems Center, Technical Report 8225,1988.
280 Bibliography
[404] A. H. Nuttall, "The Wigner distribution function with minimum spread," Naval Underwater Systems
Center, Technical Report 8317,1988.
[405] A. H. Nuttall, "Alias-free Wigner distribution function and complex ambiguity function for dis-
crete signals," Naval Underwater Systems Center, Technical Report 8533,1989.
[406] R. F. O'Connell and E. P. Wigner, "Quantum-mechanical distribution functions: Conditions for
uniqueness," Phys. Lett., vol. 83A, pp. 145-148,1981.
[407] R. F. O'Connell, "The Wigner distribution function - 50th buthday," Found. of Phys., vol 13, pp.
83-92,1983.
[408] S. Oh, R. J. Marks II, L. E. Atlas and J. A. Pitton, "Kernel Synthesis for Generalized Time-Frequency
Distributions Using the Method of Projection Onto Convex Sets," Proc. SPIE, vol. 1348, pp. 197-
207,1990.
[409] S. Oh and R. J. Marks II, "Some properties of the generalized time-frequency representation with
cone shaped kernel," IEEE Trans Sig. Proc., vol. 40, pp. 1735-1745,1992
[410) J. R. O'Hair and B. W. Suter, "Kernel Design Techniques for Alias-Free Time-Frequency Distribu-
tions," IEEE Trans. on Signal Processing, submitted, 1994.
[4111 J. R. O'Hair and B. W. Suter , "The Zak Transform and Decimated Tune-Frequency Distributions,"
IEEE Trans. on Signal Processing, submitted, 1993.
[412] J. R. O'Hair and B. W. Suter, "Multirate. A New Computational Paradigm," IEEE Trans. on Signal
Processing, submitted, 1994.
[413] J. R. O'Hair and B. W. Suter, "Kernel Design Techniques for Alias-Free Time-Frequency Distribu-
tions," Proc. IEEE ICASSP- 94, vol. pp. 333-336,1994.
1414] A. V. Oppenheim, "Speech spectrograms using the fast Fourier transform," IEEE Spectrum, vol. 7,
pp. 57-62,1970.
[4151 A V. Oppenheim and J. S. Lim, 'The importance of phase in signals' Proceedings of the IEEE, vol.
65, pp. 529-541,1981.
[416] R. S. Orr, J. M. Morris, and S. Qian, "Use of the Gabor Representation for Wigner Distribution
Crossterm Suppression," Proc. IEEE ICASSP- 92, vol. 5, pp. 29-31,1992.
[417] J. P Ovarlez, J. Bertrand, and P Bertrand, "Computation of affine time-frequency distributions
using the fast Mellin transform," Proc. IEEE ICASSP- 92, vol. 5, pp. 117-120,1992.
[418] A. Pacut, W. J. Kolodieji and A. Said, "Discrete domain Wigner distributions - a comparison and
an implementation," Proc. IEEE Int. Symposium on Circuits and Systems, pp. 1264-1267,1989.
[419] C. H. Page, "Instantaneous power spectra," Jour. Appl. Phys, vol. 23, pp. 103-106,1952
[420] A. Papandreou and G. F. Boudreaux-Bartels, "Distributions for Tune-Frequency Analysis: a Gen-
eralization of Choi-Williams and the Butterworth Distribution," Proc. IEEE ICASSP- 92, vol. 5, pp.
181-184, Man 1992.
[421] A. Papandraou, F. Hlawatsch, and G. F. Boudreaux-Bartels, "Quadratic tune-frequency distri-
butions: the new hyperbolic class and its intersection with the affine class," Sixth SSAP Workshop
on Statistical Signal and Array Processing, pp. 26-29, 1992.
[422] A. Papandreou, F. Hlawatsch, and G. F. Boudreaux-Bartels, A unified framework for the
Bertrand distribution and the Altes distribution- the new hyperbolic class of quadratic time-
frequency distributions," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp
27-30,1992.
[423] A. Papandreou, F. Hlawatsch, and G F. Boudreaux-Bartels, "The Hyperbolic Class of Quadratic
Tune-Frequency Representations - Part I: Constant--Q Warping, the Hyperbolic Paradigm, Prop-
erties, and Members," IEEE Trans. Signal Processing, vol. 41, pp. 3425-3444, Dec. 1993.
[424] A. Papandreou and G. F Boudreaux-Bartels, A Generalization of the Choi-Williams and the But-
terworth Time-Frequency Distributions," IEEE Trans. Signal Processing, vol. 41, pp. 463-472,1993
[425] J. L. Park and H. Margenau, "Simultaneous measurability in quantum theory," Intern Jour. Theoret
Phys., vol. 1, pp. 211-283,1968.
[426] T. W. Park and R. 0 Shenoy, "Time-frequency concentrated basis functions," Proc IEEE ICASSP-
90, pp. 2459-2462,1990.
Bibliography 281
[427] S. C. Pei, and T Y. Wand, "The Wigner Distribution of Linear Time-variant Systems," IEEE Trans.
Acoust., Speech, Signal Process., vol. 36, pp. 1681-1684,1988.
[428] S. C. Pei and E. J. Tsai, "Cross-terms analysis in the modified instantaneous power spectrum," IEEE
Trans. Sig. Proc., vol. 41, pp. 477-480,1993.
[429] S. Peleg and B. Friedlander, "A Technique for Estimating the Parameters of Multiple Polynomial
Phase Signals," Proc. IEEE-SP Int. Sympos Time-Frequency and Time-Scale Analysis, pp. 119-122,
1992.
[430] M. R. Perry, A Correlation Approach to Time-Frequency Representations," Proc. IEEE-SP Int. Sym-
pos. Time-Frequency and Time-Scale Analysis, pp. 249-252,1992.
[4311 A. P Petropulu, " Detection of Multiple Chirp Signals Based on a Slice of the Instantaneous Higher
Order Moments," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 261-264,
1992.
[4321 F. Peyrin and R. Frost, "A unified definition for the discrete-tame discrete-frequency, and discrete-
time/frequency Wigner distributions," IEEE Trans. Acoust., Speech, Signal Processing, vol. 34, pp.
858--867,1986.
[433] F. Peyrin, B. Karoubi, D. Morlet, P Rubel, P Desseigne, and P Touboul, "Time-Frequency and Time-
Scale Analysis of high Resolution ECGS, Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale
Analysis, pp. 77-80,1992.
[434] C. A Pickover, and L. Cohen, "A Comparison of Joint Time-Frequency Distributions for Speech,
Signals," IEEE Int. Symp. Circuits and Systems, pp. 42-45,1986.
[4351 W. J. Pielemeier and G. H. Wakefield, "Time-frequency and time-scale analysis for musical tran-
scription," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 421-424,1992.
[436] W J. Pielemeier and G. H. Wakefield, "A high resolution time-frequency representation for musical
instrument signals," J. Acoust. Soc. Am., to appear, 1994.
[4371 W. J. Pielemeier, "A Tune-Frequency Distribution with Parameter Estimators for a Musical Signal
Model," University of Michigan, Ph. D. Thesis, Ann Arbor; Michigan, 1994
[438] J. Pitton, W. Fox, L. E. Atlas, J. Luby and P Loughlin, "Range-Doppler processing with the cone-
kernel time-frequency representation," IEEE Proc. Ac. Rim Conf. on Comm., Comps. and Sig. Proc.,
pp. 799-802, 1991.
[439] J. Pitton, P Loughlin and L. E. Atlas, "Positive time-frequency distributions via maximum entropy
deconvolution of the evolutionary spectrum," Proc. IEEE ICASSP- 93, vol. 4, pp. 436-439,1993.
[440] J. Pitton, L. E. Atlas and P Loughlin, "Deconvolution for positive time-frequency distributions,"
Proc. 27th Asilomar Conf. on Sigs., Syst. and Comps., vol. 2, pp. 1450--1454,1993.
[441] J. Pitton, L. E. Atlas and P Loughlin, "Applications of positive time-frequency distributions to
speech processing," IEEE Trans. Speech and Audio Proc., to appear, 1994.
[442] M. Poletti, "Linearly swept frequency measurements, time delay spectrometry and the Wigner
distribution," J. Audio Eng. Soc., vol. 36, pp. 457-468,1988
[443] M. Poletti, "The development of a discrete transform for the Wigner distribution and ambiguity
function," J. Acoust. Soc. Am., vol. 84, pp. 238-252,1988
[444] M. Poletti, "The application of linearly swept frequency measurements," J. Acoust. Soc. Am., vol.
84, pp. 599-610,1988
[445] M. Pbletti, "Application of the Wigner-Ville distribution to time delay spectrometry," 2nd LASTED
International symposium on signal processing and its applications, pp. 900-903, 1990.
[446] M. Poletti, "Instantaneous frequency and conditional moments in the time-frequency plane,"
IEEE Trans. Signal Processing, vol. 39., pp. 755-756,1991.
[447] M. Pbletti, "The homomorphic analytic signal," 3rd LASTED International symposium on signal pro-
cessing and its applications, pp. 49-52, 1992.
[448] M. Poletti, "Linearly swept frequency measurements and the Wigner-Ville distribution," in: Time-
Frequency Signal Analysis, edited by B. Boashash, pp. 424-444, Longman Cheshire, 1992.
[449] M. Pbletti, "The development of instantaneous bandwidth via local signal expansion," Signal Pro-
cessing, vol. 31, pp. 273-281,1993.
282 Bibliography
[4501 M. Poletti, "A new description of instantaneous spectral moments," IEEE Trans Signal Processing,
submitted, 1994.
[451] M. Poletti, "The Homomorphic Analytic Signal," IEEE Trans Signal Processing, submitted, 1994.
[452] Portnoff, M. R., "Representation of Digital Signals and Systems Based on the Short-Tune Fourier
Transform," IEEE Trans. Acoust., Speech, Sig Proc., vol. 28, pp. 55-69, 1980.
1453] M. R. Portnoff, "Time-frequency representation of digital signals and systems based on short-
time Fourier analysis," IEEE Trans on Acoustics, Speech, Signal Processing, vol. 28, pp. 55-669,1980.
[454] T. E. Posch, "Wavelet Transform and Tune-Frequency Distributions" Proc. SPIE, vol. 1152, pp. 477-
482,1988.
[455] T. E. Posch, "Time-frequency distributions for a wide sense stationary random signal," IEEE Trans.
Acoust., Speech, Signal Processing, vol. 38, pp. 725-727, 1990.
[456]'E E. Posch and B R. Breed, "Space-wave number distributions and matched field processing," in.
Time-Frequency Signal Analysis, edited by B. Boashash, pp. 467187, Longman Cheshire, 1992.
[4571 1 E. Posch, "The Wave Packet Transform (WPT) as Applied to Signal Processing," Proc IEEE-SP
Int Sympos. Time-Frequency and Time-Scale Analysis, pp. 143-146,1992.
[458] R. K. Potter, G. Kopp and H. C Green, Visible Speech, Van Nostrand, 1947.
[459] D. Preis, F. Hlawatsch, P. J. Bloom, and J. A. Deer, "Wigner distribution analysis of filters with
perceptible phase distortion," Journal of the Audio Engineering Society, vol. 35, pp. 1004-1012,1987.
[460] R. Priebe and G. Wilson, Application of matched wavelets to identification of metallic transients,"
Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 349-352, 1992
[461] M. B. Priestley, "Evolutionary spectra and nonstationary processes," J. Roy. Statist. Soc. Ser. B, vol.
27,204-237,1%5.
[462] M. B. Pnestley, "Evolutionary spectra and non-stationary processes,"]. Roy. Statist. Soc., Ser. B, vol
30, pp. 1-30,1968
[463] M. B. Priestley, Spectral Analysis and Time Series, Academic Press, 1981,
[464] M. B. Priestley, Non-Linear and Non-Stationary Time Series Analysis, Academic Press, 1988.
1465] S. Puechmorel and B. Lacaze, "Optunal transforms related to paths in the time-frequency plane'
Proc. IEEE-SP Int. Sympos Time-Frequency and Time-Scale Analysis, pp. 34-37,1992.
[466] S. Qian, and J. M. Morris, "A Fast Algorithm for Real Joint Time-Frequency Transformations of
Time-Varying Signals," Electronic Letters, vol. 26, pp. 537-539,1990.
[467] S. Qian, and J. M. Morris, A New Approach for Signal Synthesis from Smoothed Versions of the
Wigner Distribution," Proc IEEE ICASSP- 91, pp. 3217-3219,1991.
[468] S. Qian and J. Morris, "Wigner-Ville Distribution and Cross-Term Deleted Representation," Signal
Processing, vol. 25, pp. 125-144,1992.
[469] S. Qian and D. Chen, "Signal Representation in Adaptive Gaussian Functions and Adaptive Spec-
trogram," Proceeding of 27h Annual Conference on Information Science and Systems, pp. 59-65,1993.
[470] S. Qian and D. Chen, "Decomposition of the Wigner-Ville Distribution," IEEE Trans. on Signal
Processing, to appear, 1994.
[471] A A. Rakes, C. Hester, R. Inguva, and B. V. K. V. Kumar, "Wigner distribution function/ Ambiguity
function processor," Proc. SPIE, vol. 936, pp 260-269,1988.
1472] P A. Ramamoorthy, V. K. Iyer, Y. and Ploysongsang, 'Autoregressive modeling of the Wigner spec-
trum," Proc IEEE ICASSP- 87, vol. 3, pp. 1509-1512,1987.
[473] P Rao, C. Griffin, and F. J. Taylor, "Time-delay estimation using the Wigner distribution," in. Pro-
ceedings of the 22nd Asilomar Conference on Signals and Systems, 1988.
[474] P Rao and F. J Taylor, "Estimation of instantaneous frequency using the Wigner distribution,"
Electronics Letters, vol. 26, pp. 246-248,1990.
[475] S. Raz, "Synthesis of signals from Wigner distributions- Representation on biorthogonal bases,"
Signal Processing, vol. 20, pp. 303-314,1990.
[476] T Reed, and H Wechsler, "Texture segmentation via non-stationanty detection using joint
spatial/spatial-frequency representations Digital Signal Processing," Proceedings of the 5th Interna-
tional Conference on Digital Signal Processing, ed. V. Capellini and A. G Constantuiides, pp. 896-900,
Elsevier, 1987
Bibliography 283
[477] T. Reed and H. Wechsler, "Tracking on nonstationarities for texture fields," Signal Processing, vol.
14, pp. 95-102,1988.
[478] K. S. Riedel, "Optimal Databased Complex Demodulation for Evolutionary Spectra," Proc. IEEE-
SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 273-276, 1992.
[479] K. S. Riedel, "Data-based kernel estimation of evolutionary spectra," IEEE Trans. on Signal Pro-
cessing vol. 41, pp. 2439-2447,1993.
[480] K. S. Riedel and A. Sidorenko, "Data Adaptive Kernel Smoothers," Computers in Physics, to appear,
1994.
[481] K. S. Riedel and A Sidorenko, "Minimum bias multiple taper spectral estimation," IEEE Trans.
on Signal Processing, to appear, 1994.
[482] K S. Riedel, "Kernel Estimation of the Instantaneous Frequency," IEEE Trans. on Signal Processing,
to appear, 1994.
[483] A. W. Rihaczek, "Hilbert transforms and the complex representation of real signals," Proceedings of
the IEEE, vol. 54, pp. 434-435,1966; and Reply by Bedrosian.
[484] A. W. Rihaczek, "Signal energy distribution in time and frequency," IEEE Trans. Info. Theory vol. 14,
pp. 369-374, 1968.
[485] A. W. Rihaczelc, Principles of High Resolution Radar, McGraw-Hill, 1968.
[486] M. D. Riley, "Beyond quasi-stationarity: Designing time-frequency representations for speech
signals,' Proc. IEEE ICASSP- 87, pp. 657-660,1987
[487] M. D. Riley, Speech time-frequency Representations, Kluwer,1989.
[488] O. Rioul, "Wigner-Ville Representations of Signal Adapted to Shifts and Dilation," AT&T Bell
Laboratories Tech. Memo., 1988.
[489] O. Rioul and M. Vetterli, "Wavelets and Signal Processing," IEEE Signal Processing Mag., vol. 39,
1564-1574,1991.
[490] O. Rioul and P. Flandrin, "Time-scale energy distributions: a general class extending wavelet
transforms," IEEE Trans. on Signal Processing, vol. 40, pp. 1746-1757,1992.
[491] G. Rizzoni and X. C. Chen, "Detection of Internal Combustion Engine Knock Using Time-
Frequency Distributions," 36th. Midwest Symposium on Circuits and Systems, vol. 1 pp. 360-363,1993.
[492] R Rohrbaugh, "Application of time-frequency analysis to machinery condition assessment, Proc.
27th Asilomar Conf. on Sigs., Syst. and Comps., vol. 2, pp. 1455-1458,1993.
[493] R. Rohrbaugh, "Advanced Tune-Frequency Analyis: Theory and Application to Machinery Con-
dition Assessment," Naval Surface Warfare Center, Technical report SADP-U93/00039-715,1993.
[494] C. E. Rosenthal and L. Cohen, "Time-Frequency Analysis of Acoustic Scattering," Proc. 24th Asilo-
mar Conference on Signals, Systems, and Computers, pp. 1061-1067, 1991.
[495] A. Royer, "Wigner Function as Expectation Value of a Parity Operator," Phys. Rev. A, vol. 15, pp.
449-450,1977.
[496] G J. Ruggeri, "On phase space description of quantum mechanics," Prog. Theoret. Phys vol. 46, pp.
1703-1712, 1971.
[497] B. E. A Saleh and N. S. Subotic, "Tune-variant filtering of signals in the mixed time frequency
domain," IEEE Trans Acoust., Speech, Signal Processing, vol. 33, pp. 1479-1485,1985.
[498] B. E. A. Saleh and M. K. An, "Filtering in the time-frequency domain," Proc SPIE, vol. 1152, pp.
426-436, 1989.
[499] B Samuny and G. Rizzoni, "Tune-Frequency Analysis for Improved Detection of Internal Com-
bustion Engine Knock," Proc IEEE-SP Int Sympos. Time-Frequency and Time-Scale Analysis, to ap-
pear, 1994.
[500] B. Samimy and G. Rizzoni, "Experimental Studies of Internal Combustion Engine Knock Detec-
tion," SAE 1995 Internationl Congress and Exposition, to appear, 1995.
[501] A. M Sayeed and D. L. Jones, "Optimal Kernels For Non-Stationary Spectral Estimation," Trans.
on Signal Processing, to appear, 1994.
[502] M. R. Schroeder and B. S. Atal, "Generalized short-tune power spectra and autocorrelation func-
tions," J. Acoust. Soc. Am., vol. 34, pp. 1679-1683,1%2.
284 Bibliography
[503] B. Schweizer and A. Sklar, "Probability distributions with given margins: Note on a paper by Finch
and Groblicld," " Found of Phys., vol. 16, pp. 1061-1064,1986.
[504) R. J. Sclabassi, M. Sun, D. N. Kneger, P Jasiukaitis and M. S. Scher, "Tune-Frequency Domain
Problems in the Neurosciences," in: Time-Frequency Signal Analysis, edited by B. Boashash, pp.
498-519, Longman Cheshire, 1992.
[505] M. Scully and L. Cohen, "Quasi-Probability Distributions for Arbitrary Operators" in: The Physics
of Phase Space, edited by Y. S. Kim and W. W. Zachary, Springer Verlag,1987.
[506] J. P. Sessarego, J. Sageloli, P. Flandrin, and M. Zakharia, "Time-Frequency Analysis of Signals Re-
lated to Scattering Problems in Acoustics," in Wavelets: Time-Frequency Methods and Phase Space, J.
M. Combes, A. Grossman, and Ph. Tchanutchian, Eds., pp. 147-153, Springer Verlag,1989.
[507] S. I. Shah, L. F Chaparro and A. S. Kayhan, "Evolutionary Maximum Entropy Spectral Analysis,"
in: Proc. IEEE ICASSP- 94, 1994.
[508] J. Shekel, "Instantaneous frequency," Proc. IRE., vol. 41, pp. 548,1953.
[509] R. G Shenoy and Z W. Parks, "Affine Wigner distributions," Proc. IEEE ICASSP- 92, vol. 5, pp
185-188,1992.
[510] R. G. Shenoy and I W. Parks, "Wideband ambiguity Functions and the Affine Wigner Distribu-
tions," Signal Processing, to appear, 1993.
[511] R. G. Shenoy and I W. Parks, "The Weyl Correspondence and time-frequency analysis," IEEE
Trans on Signal Processing, vol. 42, p. 318-331, 1994.
[512] H. Shevrin, W J. Williams, R. E. Marshall, R. K Hertel, J. A. Bond and L. A. Brakel, "Event-Related
Potential Indicators of the Dynamic Unconscious," Consciousness and Cognitum, vol.1, pp. 340-366,
1992.
[513] Y. S. Shin, J. J. Jeon, and S. G. Spooner, "Pseudo Wignerr-Ville Distribution and its Application to
Machinery Condition Monitoring," Proceedings of the 47th Meeting of the Mechanical Failures
Prevention Group," pp. 235-250,1993
[514] M. I. Skolnik, Introduction to Radar Systems, Mcgraw-Hill Book Co. 1980.
[515] D. Slepian and H. O. Pollack, "Prolate spheroidal wave functions, Fourier analysis and uncer-
tainty," Bell System Technical Journal, vol. 40, pp. 43-64,1%1.
[516] R. Smits, Accuracy of quasy-stationary analysis of highly dynamic speech signals," J. Acoust. Soc.
Am., to appear, 1994.
[517] F Soto and P Claverie, "Some properties of the smoothed Wigner function," Physica, vol.109A, pp.
193-207,1981.
[518] F. Soto and P. Claverie, When is the Wigner distribution function of multidimensional systems
nonnegative?," Jour. Math. Phys., vol. 24, pp. 97-100,1983.
[519] M. D. Srinivas and E. Wolf, "Some nonclassical features of phase-space representations of quan-
tum mechanics," Phys. Rev. D, vol. 11, pp 1477-1485,1975.
[520) L. Stankovic and S. Stankovic, "Wigner distribution of noisy signals," IEEE Trans. Sig. Proc., vol.
41, pp. 956-960,1993.
[521] L. Stankovic and S. Stankovic, "An analysis of instantaneous frequency representation using time
frequency distributions-generalized Wigner distribution," IEEE Trans. Sig. Proc., to appear, 1994.
[522] L. Stankovic, "A method for time-frequency analysis," IEEE Trans. Sig Proc., vol. 42, pp. 225-229,
1994.
[523] S. Stem and H. H. Szu, "Two dimensional optical processing of one-dimensional acoustic data,"
Opt. Eng., vol. 21, pp. 804-813,1982.
[524] L. R. O. Storey, "An investigation of whistling atmospherics," Phil. Trans. Roy Soc. vol A246, pp.
113-141,1953.
[525] H. C Strifors, C. G. Gaunaurd, B. Brusmark, and S. Abrahamsson, Aspect dependence of impulse
radar extracted signature features in the combined time-frequency domain," Proc. SPIE, vol. 2234,
1994.
[526] H. C. Strifors and C. G. Gaunaurd "Scattering of short, ultra-wideband electromagnetic pulses
by simple targets having reduced radar cross section" in: Proceed. 2nd Confer. on Ultra-wideband,
short-pulse, electromagnetic, Plenum, 1994.
Bibliography 285
[552) D. E. Vakman and L. A Vainshtein, Frequency Separation in Theory of Oscilations and Waves, Nauka
Press, Moscow, USSR, 1983.
[553] D. E. Vakman, "Computer Measuring of Frequency Stability and the Analytic Signal," IEEE Trans
Instrumentaion and Measurments, vol. 43, pp. 668-671,1994.
[554] D. E. Vakman, "Dynamic Flicker Frequency Modualtion and Noise Instabilty of Oscilators," IEEE
Trans. Circuits and Systems, vol. 41, pp. 321-325,1994.
[555] B. Van der Pol, "The fundamental principles of frequency modulation," ]our. lEE (London), vol. 93,
pp. 153-158,1946.
[556] E. F. Velez and R. G. Absher, "Transient analysis of speech signals using the Wigner time-frequency
representation," Proc. IEEE ICASSP- 89, pp. 2242-22,45,1989.
[557] E. F. Velez and R. G. Absher; "Segmentation and classification of phonemes in continuous speech
based on their time-frequency representations," Proc. SPIE, vol. 1348, pp. 188-1%,1990.
[558] E. F. Velez and R. G. Abshez "Spectral estimation based on the Wigner-Vile representation," Signal
Processing, vol. 20, pp. 325-346,1990.
[559] E. E Velez, "Speech analysis using the smoothed Wigner-Ville distribution," in: Torte-Frequency
Signal Analysis, edited by B. Boashash, pp. 351-374, Longman Cheshire, 1992.
[560] M. Vetterli, "Filter Banks Allowing Perfect Reconstruction," Signal Proc., vol. 10, pp. 219-244,1986.
[561] J. Ville, "Theorie et applications de la notion de signal analytique," Cables et Transmissions, vol.
2A, pp. 61-74,1948. Translated from French by I. Selin, "Theory and applications of the notion of
complex signal," RAND Corporation Technic al Report T-92, Santa Monica, CA, 1958.
[562] D. J. Waldo, and P R. Chitrapu, "Crossterm Reduction by a Nonlinear Combination of Time-
Frequency Signal Representations," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-ScaleAnal-
ysis, pp. 253-256,1992.
[563] W. J. Wang, and P D. McFadden, "Early detection of Gear Failure by Vibration Analysis - I. Cal-
culation of the Time Frequency Distribution", Mechanical Systems and Signal Processing vol. 7, pp.
193-203,1993.
[564] W. A. Watkins, "The Hormone Interval: Fact or Artifact in Spectral Analyis of Pulse Trains,' Marine
Bio-acoustics, vol. 2, pp. 15-43,1966.
[565] J. B. Weaver, J. Lu, and D. M. Healy, Jr. "Signal Recovery and Wavelet Reproducing Kernels," Proc
IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 177-180,1992.
[566] L. G. Weiss and R. K Young, "Wideband, Nearfield Array Processing and Wavelet Transforms,'
Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis; pp. 521-524,1992.
[567] L. G. Weiss, "Wavelets and Widband Correlation Processing," IEEE Signal Processing Magazine, vol.
11, pp.13-32,1994.
[568] Weyl, H., The Theory of Groups and Quantum Mechanics, E. P Dutton and Co., 1931. (reprinted by
Dover Publications, 1950.)
[569] A D. Whalen, Detection of Signals in Noise, Academic Press, 1971
[570] L. B. White, B. Boashash, "Time-Frequency Analysis of Non-Stationary Signals," IREECON 87,
Sydney 1987, Institution of Engineers Press.
[571] L. B. White, B. Boashash, "Pattern Classification in the Tune-Frequency Domain," Conference on
Computer Systems and Information Technology, Brisbane, 1987, Institution of Engineers Press, 1987.
[572) L. B. White, B. Boashash, "Tune-Frequency Coherence - A Theoretical Basis for the Cross Spectral
Analysis of Non-Stationary Signals," Signal Processing : Theories, Implementations and Applications,
pp. 18-23, Ed. B. Boashash, Institution of Engineers Press, 1987.
[573] L. B. White and B. Boashash, A design methodology for the optimal estimation of the spectral
characteristics of non-stationary random signals," ASSP Workshop on Spectral Estimation and Mod-
eling, pp. 65-70,1988.
[574] L. B. White and B. Boashash, "On estimating the instantaneous frequency of a Gaussian random
signal by the use of the Wigner-Vrlle distribution," IEEE Trans. on Acoustics, Speech, Signal Process-
ing, vol. 36, pp. 417-420,1988.
[575] L. B. White, "A Fast Algorithm for Pattern Classification in the Tune-Frequency Domain," in: Aus-
tralian Symposium on Signal Processing and Applications, 1989.
Bibliography 287
[576] L. B. White, "Resolution Enhancement in Time-Frequency Signal Processing using Inverse Meth-
ods," Proc. IEEE ICASSP- 89, pp. 1622-1625,1989.
[577] L. B. White, "Tune-Frequency Filtering from Convex Projections," Proc. SPIE, vol. 1348, pp. 158-
169,1990.
[578] L. B. White, B. Boashash, "Cross Spectral Analysis of Nonstationary Processes," IEEE Trans Infer
Theory, vol. 36, pp. 830-835,1990.
[579] L. B. White, "Estimation of the instantaneous frequency of a noisy signal," in: Time-Frequency
Signal Analysis, edited by B. Boashash, pp. 233-249, Longman Cheshire, 1992.
[580] L. B. White, "Signal Synthesis from Cohen's Class of Bilinear Time-Frequency Distributions using
Convex Projections," Proc. IEEE ICASSP- 91,1991.
[581] L. B. White, "Transition Kernels for Bilinear Time-Frequency Signal Representations," IEEE Trans.
Signal Processing, vol. 39, pp 542-544,1991.
[582] L. B. White, "The Wideband Ambiguity Function and Altes' Q-distribution Constrained Synthesis
and Tune-Scale Filtering," IEEE Trans. Information Theory, vol. 38, pp. 886-892,1992.
[583] S. E. Widmalm, W. J. Williams and C. Zheng. "Time Frequency Distributions of TMJ Sounds," Jour-
nal of Oral Rehabilitation, vol. 18, pp. 403-412,1991.
[584] E. P. Wignei. "On the quantum correction for thermodynamic equilibrium," Physical Review, vol.
40, pp. 749-759, 1932.
[585] E. P Wignei; "Quantum-mechanical distribution functions revisited," Perspectives in Quantum The-
ory, W. Yourgrau and A. van der Merwe, eds., pp. 25-36, MIT Press, 1971.
[586] J. E. Wilbur, and F. J. 'Taylor, "High-speed Wigner processing based on a single modulus quadratic
residue numbering system," Proc. IEEE ICASSP- 87, pp. 1031-1034,1987.
[5871 J. E. Wilbur, and F. J. Taylor, "An acoustooptic Wigner processor for time-varying signal analysis,"
Proceedings of the IEEE, vol. 75, pp. 427-428,1987.
[588] J. E. Wilbur, and F J. Taylor, "Consistent speaker identification via Wigner smoothing techniques,"
Proc. IEEE ICASSP- 88, vol. 1, pp. 591-594,1988.
[5891 It M. Wilcox, "Exponential Operators and Parameter Differentiation in Quantum Physics," Jour.
Math. Phys., vol. 8, pp. 962-982,1%7.
[590] W. J. Williams, H. Shevrin and R. E. Marshall, "Information Modeling and Analysis of Event Re-
lated Potentials," IEEE Tins. on Biomedical Engr., vol. 34, pp. 928-937,1987.
[591] W. J. Williams and J. Jeong. "New Tune-Frequency Distributions for the Analysis of Multicompo-
nent Signals," Proc. SPIE, vol. 1152, pp. 483-495,1989.
[5921 W. J. Williams, J. Jeong, M. Driscol, and S. Krishnamachari, Applications and Interpretations of
Reduced Interference Time-Frequency Distributions," in: 24th Asilomar Conf on Signals, Systems,
& Computers, vol. 2,1049-1054,1990.
[593] W. J. Williams, M. L. Brown, and A. O. Hero M. "Uncertainty, Information and Time-Frequency
Distributions," Proc. SPIE, vol. 1566, pp. 144-156,1991.
[594] W. Williams and J. Jeong, "Reduced interference time--frequency distributions," in: Time- Frequency
Signal Analysis, edited by B. Boashash, pp. 74-97, Longman Cheshire, 1992.
[5951 W. J. Williams. "Tune-Frequency Analysis of Biological Signals" Proc. IEEE ICASSP- 93, vol. 1, pp.
83-%,1993.
[5%] P J. Wiseman and G. F. Boudreaux-Bartels, "Wigner Distribution Analysis of Acoustic Well Logs,"
Proc. IEEE ICASSP- 87, pp. 2237-2240,1987.
[5971 K M. Wong and Q. Jin, "Estimation of the time-varying, frequency of a signal: The Cramer-
Rao bound and the application of Wigner distribution," IEEE Trans. on Acoustics, Speech Signal
Processing, vol. 38, pp. 519-536,1990.
[598] J. C. Wood, and D. I Barry, "Radon transformation of time-frequency distributions for analysis of
multicomponent signals,' Proc. IEEE ICASSP- 92, vol. 4, pp. 257-261,1992.
[5991 J. C. Wood, and D. I Barry, "Radon transform of the Wigner Distribution," Proc SPIE, vol 1770,
pp. 358-375,1992.
[600] J. C. Wood, A. J. Buda, and D. I Barry, "Tune-frequency transforms. A new approach to first heart
sound frequency analysis," IEEE Trans in Biomedical Engineering, vol. 39, pp 728-739,1992.
288 Bibliography
[60] ] J C. Wood, and D. T. Barry, "The variation of first heart sound frequency dynamics across the
human chest wall," Computers in Engineering in Medicine and Biology, in press, 1994.
[602] J. C. Wood, M. P Festen, M. J. Lim, J. Buda, and D. T Barry, "Differential effects of myocardial
ischemia on regional first heart sound frequency," Journal of Applied Physiology, to appear, 1994.
[603] J. C. Wood, and D. T. Barry, "Radon transformation of time-frequency distributions for analysis of
multicomponent signals," IEEE Trans. of Acoustics, Speech, and Signal Processing, in press, 1994.
[604] J. C. Wood, and D. T Barry, "Time-frequency analysis of the first heart sound: techniques and
applications," Engineering in Medicine and Biology Magazine, submitted, 1994.
[605] J. C. Wood, and D. T. Barry, "Analysis of multicomponent, nonstationary biological signals" Medical
and Life Sciences in Engineering, in press, 1994.
[606] J. C. Wood, and D. T Barry, "Linear signal synthesis from the Radon-Wigner domain," IEEE Trans.
of Acoustics, Speech, and Signal Processing, vol. 42, 2094-2104,1994.
[607] J. C. Wood, and D. T Barry, "Tomographic reconstruction of the time-frequency plane," IEEE Trans.
of Acoustics, Speech, and Signal Processing, vol. 42, 2105-2111, 1994
[608] J. C. Wood, "Frequency dynanucs of the first heart sound," Ph. D. thesis, University of Michigan,
Ann Arbor, M1, 1994.
[609] P. M. Woodward, Probability and Information Theory with Applications to Radar, Pergamon Press,
1953.
[610] H. Xie and J. M. Morris, "Design of Orthogonal Wavelets with Better Time-Frequency Resolution,"
Proc. SPIE, vol. 2242, pp. 878-887,1994.
[611] N. Yen and L. R. Dragonette, "Inverse acoustic scattering problem: time and frequency approach,"
Review of progress in quantitative nondestructive evaluation, Ed. by D. O. Thompson and D. E. Chi-
menti, vol.10A, pp. 89-96, Plenum, 1981
[612] N. Yen, "Time and frequency representation of acoustic signals by means of the Wigner distribu-
tion function," J. Acoust. Soc. Am., vol. 81, pp. 1841-1850,1987.
[613] N. Yen, L. R. Dragonette and S. K Numrich, "Graphical representations of the spherical-shell-
forin-function based on Wigner distribution algorithm," J. Acoust. Soc. Am., vol. 84, suppL 1, pp.
S185, Fall 1988.
[614] N. Yen, L. R. Dragonette, and S. K Numrich, "Tune-frequency analysis of acoustic scattering from
elastic objects," J. Acoust. Soc. Am., vol. 87, pp. 2359-2370,1990.
[615] N. Yen and L. R. Drogonette, "Application of time-frequency analysis to acoustic response from an
insonified object," in: LASTED, Signal Processing and Digital Filtering, Ed. by M. H. Hamza, ACTA
press, 1990.
[616] N. Yen, "Wave packet decomposition," J. Acoust. Soc. Am., vol 95, pp. 889-896,1994.
[617] R. K Young, Wavelet Theory and Its Applications, Kluwer Academic Publishers, 1993.
[618] K B. Yu and S. Cheng, "Signal synthesis from Wigner distribution," Proc IEEE ICASSP- 85, pp.
1037-1040,1985.
[619] K B. Yu and S. Cheng, "Signal synthesis from pseudo-Wigner distribution and applications," IEEE
Trans. Acoust., Speech, Signal Processing, vol. 35, pp. 1289-1302,1987
[620] K B. Yu, "Signal representation and processing in the mixed time-frequency domain," Proc. IEEE
ICASSP- 87, pp. 1513-1516,1987.
[621] L. A. Zadeh, "Frequency analysis of variable networks," Proc. IRE., vol. 38, pp. 291,1950.
[622] E. J. Zalubas and M. G. Amin, "The design of signal dependent time- frequency kernels by the
McCllelan Transformation," Proc. IEEE-SP Int Sympos. Time-Frequency and Time-Scale Analysis, pp.
91-94,1992.
[623] H. P Zaveri, W. J. Williams, L. D. Iasemidis and J. C. Sackellares. "Time-Frequency Representation
of Electrocorticograms in Temporal Lobe Epilepsy." IEEE Trans on Biomed. Engr., vol. 39, pp. 502-
509,1992
[624] B. Zhang and S. Sato, "A New Time-Frequency Distribution and Its Application to Speech, Signal
Processing," Proc. IEEE-SP Int. Sympos. Time-Frequency and Time-Scale Analysis, pp. 417-420, 1992.
[625] B. Zhang and S. Sato, "A time-frequency distribution of Cohen's class with a compound kernel
and its application to speech signal processing," IEEE Trans Sig Proc, vol. 42, pp. 54-64,1994.
Bibliography 289
[626] Y. Zhao, L. E. Atlas, and R. J. Marks, "The use of cone-shaped kernels for generalized time-
frequency representations of nonstationary signals," IEEE Trans. Acoust., Speech, Signal Processing,
vol. 38, pp. 1084-1091,1990.
[627] C. Zhen, S. E. Widmalm and W. J. Williams, "New time-frequency analyses of EMG and TMJ sound
signals," Proc. IEEE Int. Conf. Engineering in Medicine and Biology, pp. 741-742,1989.
[628] C. S. Zheng, W. J. Williams and J. C. Sackellares, "RID Time-Frequency Analysis of Median Fil-
ters and Lowpass Filters in Reducing EMG Artifacts in EEG Recording," IEEE Intl. Conference on
Engineering in Medicine and Biology, vol. 1, pp. 350-351,1993.
[629] Y. M. Zhu, F. Peyrin and R. Goutte, "Equivalence between two-dimensional analytic and real signal
Wigner distributions," IEEE Trans. Acoust., Speech, Signal Processing, vol. 37, pp. 1631-1634,1989.
[630] Y. M. Zhu, F. Peyrin and R. Goutte, "The use of a two dimensional Hilbert transform for Wigner
analysis of two-dimensional real signals," Signal Processing, vol. 19, pp. 205-220,1990.
[631] Y. M. Zhu, Y. Gao, and R. Goutte, "An Investigation of the Wigner Model for Texture Discrimina-
tion," Proc. IEEE-SP Int. Sympos. Time-Frequency and rime-Scale Analysis, pp. 539-542,1992.
[632] T P. Zielenski, "New detection method based on the cross-terms mechanism of the Wigner-V lle
transform," Signal Processing V. - Theories and Applications, Eds. L. Torres, E. Masgrau and M. A.
Lagunas, Elsevier, vol. 1, pp. 157-160,1990.
Index
A Average value, 4, 5, 8-20, 54-56
Ackroyd, M. H., 45 and analytic signal, 28-29
Adjoint, see operator conditional, 63-64
Altes, R. A., 256 of an operator, 213-215
Ambiguity function, 96,108,116-117,125, of frequency, see frequency
137,142,166 of Hermitian operators, 214
generalized, 137 of non-Hermitian operators, 215
Amin, M. G., 189,193 of scale, see scale
Amplitude of time, see time
definition of, 3, 29, 35-39
determination of, 3, 35-39
modulation, 3,18-21,38,110,121,126, B
203 Bandlimited, 33-36, 40, 87, 115, 129, 144,
see also instantaneous frequency, an- 194
alytic signal Bandwidth equation, 15--17,179,223,224
Analytic signal, 28-42,130,196, 201 Bandwidth, 8-19,46-52
and the complex signal, 27-28 AM and FM contribution to, 17-19
calculation of, 31-34 expressed in terms of signal param-
energy of, 31 eters, 15-17
factorization theorem, 34-35 instantaneous, see instantaneous band-
Hilbert transform and, 30-31 width
imaginary part of, 30 Baraniuk, R. G., 166,167, 235,256
interpretation of, 35-36 Baraniuk-Jones optimal kernel design, 166--
of a complex signal, 31 167
of an analytic signal, 32 Barry, D. T,135,171
of derivative, 32 Basis functions, 204-205
of the convolution of two signals, 33 Bedrosian, E., 35
of the product of two signals, 34-36 Bernard, G., 78, 203
of the sum of two signals, 34 Bertrand, J., 256
paradoxes, 40 Bertrand, P., 256
quadrature approximation, 36-39 Bescos, J., 184
Anticommutator, 44, 209, 212 Bilinearity, 116,140
Antinormal ordering, see correspondence Born, M., 44,174
rules Born-Jordan distribution, 113, 138, 158-
Atlas, L. E., 78, 138, 144, 162, 166, 172-175, 160,174
201-203 Born-Jordan rule, see correspondence rule
Autocorrelation function, 23,193 Boudreaux-Bartels, G. F., 165,191, 256
local, 137, 230, 258 Bunsen, R. W, 2
291
292 Index
time-frequency, 84-85
C Conditional density, 63-64
Calculation tricks, 8-15 time-frequency, 84-85
Car door slams, 77 Conditional standard deviation, 64-65,85,
Cartwright, N., 132 90,103,134,178-179
Cauchy-Riemann conditions, 30 Condon, E. U., 44
Chaparro, L. F., 193 Cone kernel, see Zhao-Atlas-Marks distri-
Characteristic function operator bution
and moment method, 158-161 Correlation coefficient, see covariance
and scale 256, 257, 260, 261 Correspondence rules
averages,157 and characteristic function, 152-156
evaluation, 154-156 and distribution, 152-157
for arbitrary variables, 225-229 antinormal, 158
for general class, 156-157 Born-Jordan, 153,159
for time and frequency, 152-161 general (Cohen), 153,159
for time and scale, 256 normal, 153,158
Characteristic function, 56-69, 83, 91, 96, symmetrization,153,158
113,116-118 Weyl,153,158
for spectrogram, 96 Covariance, 63
for the frequency density, 22-23
and characteristic function, 63
for the general class, 137
and correlation coefficient, 63
for the time density, 22-23
and uncertainty principle, 46-48,216-
for Wigner distribution, 116
218
one dimensional, 56-59
for arbitrary quantities, 216-217, 233
relation to moments, 58
for time and frequency, 20-23, 215
two dimensional, 59-65
when is a function a, 58 in terms of the spectrum, 21
of a real signal, 22
Chen, D., 187
of a real spectrum, 22
Choi, H. I., 162
Choi-Williams distribution, 138 144, 147, of a signal, 20-22
164-165, 168-172 Cristobal, G., 184
Choi-Williams method, 168-172 Cross terms, 104, 116, 125--131, 141, 163,
Claasen, TA.C.M., 114,137, 201 166,169,172-175,187,192
Cohen, L., 138,153,159 Cross distribution, 139
Cole, N. M., 135 Cross Wigner distribution, 124, 129
Commutator Cumulative distribution function, 54
for arbitrary operators, 209, 210 Cunningham, G. S., 162, 168, 188, 194
for time and frequency, 211
involving scale, 244
Completeness, 244,253 D
Complex distributions, 189 Darwin, C. G., 44
Complex energy spectrum, 174-176 Decay rate model, 180
Complex signal, 3,27-43 and formant bandwidth, 180
and real signals, 124 Delta function distribution, 185-186
reasons for, 28, 29 Density function, 54, 83, 92, 96, 246
representation, 27 of a new variable, 68, 218, 219
see also analytic signal of a single variable, 219-224
Compression operator, 242,243, 250 of instantaneous frequency, 41
Conditional average, 64, 134 one dimensional, 53-56
Index 293
H
Heart sound, 80,171 K
Heat equation, 2 Kayhan, A. S., 193
Heisenberg, W, 44 Kernel design, 162-167
Hermitian functions, 188 for product kernels, 165
Hermitian operator, 12-14, 204-215 for reduced interference, 162-165
averages, 213-214 optimal, 166-167
eigenvalue problem, 205-206 projections onto convex sets, 166
forming of, 209-210 Kernel method, 136-151
function of, 210-211 Kernel
importance of, 206 and covariance, 146
product of two, 212,215 and finite support, 144
range of eigenvalues, 206 and instantaneous frequency, 147-148
reality of eigenvalues, 214 and marginals, 141
Hilbert transform, 30-31 and reality, 142
Hlawatsch, E, 256 and scale invariance, 143
Hook, R., 80 and total energy, 142
Index 295
0
M O'Hair, J. R., 189,194
Margenau, H., 113 Oh, S., 166,173
Margenau- Hill distribution, 113,136,138, Operator
174-177 adjoint, 211
Marginal, 59-64 algebra, 209
and time-frequency distributions, 83 eigenfunctions of, 205
for arbitrary variable, 219, 225 eigenvalues of, 205,214
frequency, 82-83 commutator
inverse frequency, 238-239 for arbitrary operators, 209-210
scale, 246, 255 for time and frequency, 211
time, 82-83 involving scale, 244
Marinovich, N. M., 191, 256 frequency, 11-15
Mark, WD.,114 functions of, 210
Marks II, R. J., 138, 172, 173, 162, 166 Hermitian, 12-14, 204-215
Martin, W, 193 averages, 213-214
Mecklenbrauker, WF.G.,114,137,201 eigenvalue problem, 205-206
Mellin transform, 246 forming of, 209-210
Minimum uncertainty product signals function of, 210-211
for arbitrary variables, 217-218 importance of, 206
for time and frequency, 48-49 product of two, 212, 215
for time and scale, 251-252 range of eigenvalues, 206
Modified signal, 94-95 reality of eigenvalues, 214
Moment operator method, 158-161 inverse, 211
Moments, 55-63 linear, 211
296 Index