CHAPTER-6

EXTENDED PROTOCOL FOR HIGH-EFFICIENCY VIDEO CODING

TO ENHANCE PERCEPTUAL QUALITY OF REAL-TIME
STREAMING VIDEO FRAMES

The previous chapter mostly concerned about formulating an analytical methodology

includes incorporation of NTSC standard for processing high quality streaming video frames.
However, the prime of focus of the chapter lies into conceptualizing three different levels of
encoding schemes integrated with H.265 standard along with a MLO mechanism which
initiates a novel concept of orthographic projections for extracting features considering
current and neighbouring pixel information. It further aims to perform a high-end
compression while retaining high-quality visual impact into the received frames. Moreover,
the enhancement in technological aspects broadly gives an insight into the possibility which
concerns the possibility of implementing a much more advanced methodology for the
enhancement of perceptual quality of a video frame. This chapter carries out the extensive
study aims to design an analytical methodology pertain to enhance perceptual quality and
resolution of an encoded video frame (i.e. from a streaming video frame sequence). It
implicates the continuation of the previous work by enhancing HEVC encoding to a certain
extent.

6.1 INTRODUCTION

Video compression codec usually belong to the family of streaming applications,

which are basically a part of collaborative communications. However, in the last two decades
witnessed a widespread multi-directional technological development in the field of
collaborative communication networks such as telecommunications, networking, digital
image/video processing, etc. Hence, video compression or video encoding extended its scope
of applicability into various fields such as online streaming of video frames, MPEG data
compression over wireless networks etc. The reason behind applicability of video coding to a
possible extent discloses the fact that supports its importance on reducing the amount of data
required to represent a digital video signal prior to transmission or broadcasting. As in the
current scenario of various technological aspects, real-time data communications suffer a lot
from storage and bandwidth constraints, thus keeping that into the mind researchers focused

89
on designing efficient video compression techniques. The extensive analysis to get insight
into the conventional video codec shows that the prime perspective of the video compression
lies into achieving optimal trade-off between Superior Quality Data and Higher
Compression Ratio. It also shows that in the recent times more emphasis has been given into
H.265 for its higher performance and less design complexities. It also found to achieve
significant outcomes in the context of video quality resolution along with high end
compression as compared to its legacy versions (i.e. H.263/4). The proposed framework in
this manuscript adopts H.265 over the HEVC based video compression because of its
compatibility with mobile devices and higher performance towards assessing higher
compression performance of 8K (UHD) videos with minimum utilization of channel capacity
[1][2]. The chapter further formulates a novel concept of HEVC protocol based on H.265 to
attain higher perceptual quality of encoded video frames (8K UHD). The findings of the
study explicitly prove the fact that proposed technique significantly achieves an optimal
trade-off between hardware compatibility with compression cum 8K resolution. It also
depicts the superiority of the proposed algorithm in terms of compression performance,
quality of the signal and redefines better resolution standards. The implementation supports
enhancing the compressed video quality while retaining the 8K UHD resolution on mobile
devices [3][4].

6.2 PROBLEM FORMULATION

This section intended to present the existing research issues identified in the
conventional studies. Moreover, it also gives an insight into the reasons behind those issues.
A closer look into the design flaws of the conventional video encoding standards such as
H.263 and H.264 shows that there are certain negligence‟s present in the existing conceptual
models which are needed to be clarified. More or less, the detailed elaboration of
conventional video compression methodologies and its impact towards utmost retention of
8K UHD videos in mobile communication highlighted in Chapter Two of the proposed thesis
from fundamental aspects. It also presents the significant research issues associated with the
state of art studies which further helps formulating the problems to be focused on much
precisely. In order to draw the attention of the future researchers to the fact concerning the
research gap concluded from the existing studies which further helps formulating much better
video compression techniques the current work explicitly reviews the conventional studies
and discusses few of the significant flaws in this context. A closer look into, majority of the

90
existing research works carried out till date shows that it only dragged down its focus
towards video compression using H.265 codec. On the other hand, very less emphasized
given on 8K retention of HD videos in the existing mobile communication scenario.
Although H.265 achieves very high-quality video compression by means of intra or inter
prediction mechanisms, which further reduces the rate-distortion factor. Still there exist some
sorts of factors that do not guarantee constancy of 8K resolution in a compressed video file
transmitting through a mobile communication channel. Therefore, some of the conventional
HEVC algorithms do not impose retention of 8K resolution while performing compression
on a video file as it requires a very high scale of resource utilization such as buffer, channel
capacity etc. The implementation of conventional HEVC or H.265 algorithms to a certain
extent makes them incompatible for mobile communication.

There exists different type of compression mechanisms pertaining various forms and
compatibility requirements for built-in functionalities of different HEVC standards. The
current studies also lack compression standards where MLO can be implemented efficiently
irrespective of any additional spaces or time complexity constraints. Therefore, the adoption
of the MLO based encoding to initiate a balance in between video compression and the 8K
retention of UHD videos plays a significant role in the proposed algorithm. It also offers a
better round of solution by reducing the cost of implementation and the utilization of
additional resources in comparison with the existing baselines. There are very less amount of
studies found to use useful performance parameters, which are needed for benchmarking of a
proposed solution towards a particular problem (e.g. in our case to enhance the perceptual
quality of encoded video frames while performing high-end compression using H,265 based
HEVC protocol). Therefore the problem statement in this context is defined as follows.

“It is difficult to design a less complex video coding technique for performing high end
compression along with enhancing the perceptual quality of UHD video frames in a mobile
network communications”.

The next segment of the manuscript formulates the study objective and its respective
solution approach by referring the above stated problem description and research issue.

91
6.3 OBJECTIVE

The objective in this context is formulated by concerning utilization of most recent

video compression protocol namely, H.265 standard for the purpose of achieving better
perceptual quality in real time streaming videos in peak network traffic conditions. The
objective is finally stated as follows.

- To increase the perceptual quality of video frames for supporting 8K (UHD) resolution.

The next section further discusses the proposed solution given in the context of
supporting the above-stated objective.

6.4 PROPOSED SOLUTION

The proposed solution is hereby stated by supporting the above mentioned objective.
It is subjected to design a novel concept of HEVC encoding based on H.265 Protocol to
ensure an optimal trade-off in between high-end compression and maintaining data quality. It
also aims to increase the perceptual quality of encoded video frames to further support 8K
(UHD) resolution at the recipient end. The proposed solution is defined by keeping into the
mind that it should poses aid of extensibility to supports the variable communication features
of future mobile networks. The present work is the extended version of our prior research
work where the recent HEVC encoding standard has been integrated with H.265 standard to
maintain a balance between compression and data quality. The present work explicitly makes
some sort of considerations to implicate the aim of the proposed research work such as

1. Keeping into the mind, the conventional design flaws of state of art compression
protocols, the proposed approach aims to design a framework suitable for increasing
the 8K resolution of UHD video frames while achieving high end compression.
2. To present a technique which can assure, zero loss of data during compression of
metadata of a unit frame.
3. Formulating an analogy which accesses the performance of the proposed framework
with respect to different parameters such as compressed bits, data quality and
algorithm processing time.

92
6.5 CONCEPTUAL MODEL

The conceptual model distinctively designed to support the objective as stated above.
It introduces an analytical model of an integrated framework of HEVC for the purpose of
enhancing perceptual quality of compressed video frames on the recipient side. However, the
proposed scheme also subjected to consider high-end compression of video frame planes by
maintaining the 8 K (UHD) resolutions where the PSNR values extended up to the range of
35-45 dB.

Communication
Connectivity Establishment

Terminal
Server Unit

No of Frames
Upload bit stream to the Network

Segment Length

Integrated HEVC Client Unit

Encoding

Installing DASH Receive encoded

Server Video frames
- IP Address
Integrated HEVC
- Port Number Decoding

View Video

Figure 6.1 Proposed Conceptual Model

Apparently, the proposed conceptual architecture highlighted above in Fig 6.1 aims to
minimize streaming video quality degradation using integrated HEVC video codec standard.
However, the proposed system introduces a multi-level optimization scheme on pixel
properties to achieve better retention of perceptual quality for High Definition (HD)
streaming video frames without producing any computation flaws. This section broadly
discusses the proposed HEVC encoding scheme followed by three different levels of
optimization procedures (i.e. Level-1, Level-2 and Level-3). The design and development of
the proposed system is subjected to perform the encoding and decoding in both server and
receiver side while the transmission of the encoded streaming video frames (e.g. High

93
Definition (HD) , Ultra High Definition (UHD) and Standard Definition (SD)videos) take
place on a Dynamic Adaptive Streaming over HTTP (DASH)Server. The DASH server
installation for communication set up efficiently connect the client by detecting its IP address
and transmit the encoded video stream to the client side by creating a socket layer. The
socket includes different headers to the data packet during the transmission mode such as
Link Header, Network Header, and Transport Header etc. which is preferable termed as UDP
Encapsulation (Extension). However, the mode of transmission of the video frames therefore
defined by the following communication scenario which includes

1) Communication establishment in between the Network and Transport layer.

2) Finding correct network interface i.e. MAC address from the Link Layer
3) Detection of correct port number enabled in Transport Layer.
4) Ordering of information from the Transport Layer (Reassembling)
5) Delivery of the intended message from the Application Layer.

The step by step procedures for implementing the proposed high efficiency video
codec standard in a TCP/IP protocol stack is discussed below.

6.5.1 Conceptualisation of Uploading Video Frames (UHD/HD/SD) in Server Unit

This section discusses the procedure of uploading a video object into the proposed
unit. Moreover, the study adopted the processing of most modern HEVC standard with H.265
Codec while implementing a multilevel optimization scheme on both server and receiver
units. The algorithm-I is thereby depicted below.

Table 6.1Notation for Algorithm-I

Sl. No. Symbolizations Description

1 Vfile Video File
2 Vl Loaded Video
3 Vid Variable to define Video Object
4 fn Video file name
5 fp Directory/file Path
6 ft File type (.avi/.mpg/.wmv)
7 Vpath Video path
8 µ(fn , fp) Function operates on storing and

94
 recreating video path
9 VR Video Reader Function
UI (ft) Function to opens up modal dialogue
10
box to display list of files.
11 f1 First frame
12 Noframes Number of frames

Algorithm-I: Import a video Object for Processing Frames

Input: Vfile
Output: Vl
Start
6. Globally define Vid
7. Define fn , fp
8. µ(fn , fp) UI (ft)
9. Vpathµ(fn , fp)
10. Construct a video object
Vid VR (Vpath)
11. Read the first frame
f1 read (Vid, 1)
12. Resize the frame by (128x128)
13. Conversion of RGB to GREY for f1
Apply f1 RGB2GREY (f1)
14. Show f1
End

The above algorithm conceptualized for enabling the video processing in the server
unit. The above algorithm states the scenario to import a video object into the proposed
framework for assessing a sequence of frame. It also shows that in the preliminary level the
framework imports a video object (i.e. Vid in Line-1) by taking Video file name, file path and
type into consideration. It process an in-built MATLAB function namely Video Reader () to
subsequently read video frame sequence from MATLAB workspace. It can also be seen that
VR supports different file formats which may have a variation with respect to different
platform dependencies such as in windows it only supports MPEG-1 (.mpg) , Windows

95
Media Video (.wmv) extensions etc. The extensibility added into the functionality shows that
VR generates a video object file to read the respective video frame sequence (Video Data) by
acquiring the individual file directory tree structure (file path, file name). The function VR
mostly gain an insight into properties of a video file by reading its name, path, Tag, User
Data, Width, Height, Frame Rate, Bits per pixel values, Format:‟RGB24‟ etc. This
preliminary stage also considers processing a sequence of video frames which can be
initiated and operated only if a video object file is created. The video object takes a video file
into consideration and converts it into frames by acquiring its pixel information. The
following algorithm aims to achieve the above stated operation irrespective of time
complexity.

This phase of the overall system also considers defining segment length and number
of video frames to be encoded before proceeding towards HEVC encoding scenario.

6.5.1.1 Experimental Measurements carried out for Algorithm-I

The following outcomes are explicitly measured by simulating the above algorithm in
a prototyped test bed where it reads each frame of a video object and convert it into a grey
scale featured frame.

 Case:1 (To read each frame of a video object of .avi extension )

The video file namely fun.avi is considered for simulating the above mentioned
algorithm I in a parameterized test bed. The outcomes of the simulation (for only 1 frame)
are highlighted in the following Table 6.2. Initially the proposed model considers importing a
video object with the size of 240x320 frame sizes and later on it converts the frame into a
greyscale version of size 128 x 128.

Table 6.2 Resizing Frame-1 (128x128)

RGB Frame (Before Size f1 (resizing) Size

Conversion)
240x320 128x128

96
 Another Table 6.3 shows the grey scale version of the resized first frame (f1) of a
video sequence after being processed through a grey scale conversion.

Table 6.3 (Conversion of frame-1 into a Grey Scale Image)

f1 (Before Conversion) Size f1 (resizing) Size

128x128 128x128

Further, the video frame sequence is encoded using the integrated version of HEVC
encoding where the number of frames Noframes and the segment length play a crucial role. The
proposed HEVC encoding parameters are precisely taken into consideration to attain high
end compression without affecting the video quality (i.e. retention of 8k resolution). The
study, therefore, aims to achieve an optimal trade-off between the compression ratio and the
high quality of the streaming video frames. The following is the integrated HEVC encoding
scheme intended to perform three different levels of optimization for retention of 8K (HD)
resolution at the receiver end. However, the performance evaluation of the proposed
algorithm justifies its significant contribution to accessing high-end compression as well as
ensuring high quality of encoded video frame in the next segment.

Table 6.4Notation for Algorithm-II

Sl. No. Symbolizations Description

9. Vfile Sample video file
10. fs Start Frame
11. fl Last Frame
12. EVframes Encoded Video Frames
13. Noframes Number of frames
14. Mf Matrix of frame
15. read() Read the video File
Computation of Monochrome
16. I
Luminance (NTSC Standard)
17. Size (f) Size of a frame
18. Psize Macro block size for P frames
19. Sl Segment Length

97
Level-1 Encoding (Optimization)

Algorithm-II: HEVC Encoding and Optimization Pipeline (Level-1)

Input: Vfile, fs, fl, i, Sl

Output: EVframes (Level-1)

Start
10. Input Vfile
11. Init fi, fl , i , Psize
12. Compute Noframesfl- fi + 1
13. Read the first frame

f1 read (Vid, 1)

14. Compute size (f)

15. Compute row & col multiple of block size

 n   nc 
nr= round  r  *bi* 4, ncround   * bi* 4
 bi / 4   bi / 4 

16. nr , nc , k  size (f)

17. Calculate Total Bits Per Frame
18. for ifs : fl
19. f  read (Vid) Read frame
20. Resize the frame by (128x128)
21. Mfread(Vfile, i)
22. Compute the monochrome luminance (Multimedia coefficient Estimation)

I Mf . (α+β+δ)

23. i i+1
24. end of for

End

The above algorithm applies the proposed HEVC encoding followed by Level-1
optimization on the pixel properties of a streaming video frame. Although the prime
objective of the above algorithm 2 is, segmentation of video frame sequence for the ease of
HEVC encoding but our proposed framework brings a significant contribution by adding up
an additional extension to extract the features of each frame (Line-5, 6, 7.8). It further resizes
each frame to extract multimedia coefficient of pixel properties as well as the monochrome
luminance (Line-13). The algorithm execution states that it works very efficiently for a

98
specific range of frame sequence and creates a matrix of frames (Line-12). However, the
multimedia coefficients are denoted with α, β, δ where the values for each multimedia factor
are (i.e. .2989, .5870, and .1140) considered respectively (Line-7). Further, the proposed
algorithm-2 increments the frame number until the processing of video frames reach up to the
last one. The level-1 optimization and encoding process partially improves the perceptual
quality of the encoded video. For better retention of resolution (i.e. obtaining higher SNR)
and compression ratio the video object is further processed through other two levels of
optimization and encoding which are discussed after the completion of this segment.

6.5.1.2 Experimental Measurements for Algorithm-II

The following Table 6.5 highlights the outcomes obtained from simulating the above
algorithm for (1-5) frames of an .avi video object. It basically shows the monochrome
luminance extraction and the computation of total number of bits per frame. The process
also encodes the video frame by incorporating HEVC Level-1.

Table- 6.5Anencodes the video frame by incorporating HEVC Level-1

Frame Index Encoded Frame (Level-1)

1
(I-frame)

99
 5

The proposed framework also incorporates (Level-2) encoding and optimization

process which is a further continuation of the prior Algorithm-II. In this context the
algorithm the prior Level-1 encoded/Optimized video sequence as an input and performs
Level-2 Optimization by extracting its features such as height, width, dimension etc. In the
Level-2 optimization, the system basically focuses on designing integrated HEVC protocol
using H.265. In the initial stages of the algorithm considers initializing encoding parameter
called quality parameters (Qp). The conceptual modelling of the proposed Algorithm-III
which performs Level-2 optimization to maintain the perceptual quality of video frames
along with maximum level of compression is given below.

Table- 6.6Notation for Algorithm-III

Sl. No. Symbolizations Description

1. Vfile Sample video file
2. fs Start Frame
3. fl Last Frame
4. Qp Quality Parameters
5. h Height
6. w Width
7. Vseq Video Sequence
8. N Number of frames
9. bits Bit stream
10. bi Block size
11. BR Bit rate
12. ET Encoding time
13. Rf Receiving frames

100
 Level-2 Encoding (Optimization)
Algorithm-III: HEVC Encoding Pipeline with H.265 (Level-2)
Input: Vseq, Qp

Output: EVframes (Level-2)

Start

1. Input Vseq
2. Init Op
3. Global h,w
4. Switch for extended Motion Estimation
5. Extract input video sequence dimension
h, w, N  size (Vseq)
6. Init PSNR for each frame
7. Init BR for each frame
8. Init reconstructed Vseq
9. Init output bit stream
10. Save header by applying
Bits header (h,w,Qp, fs,fl, Bsize)
11. Save Bits into bits
12. Encode I-frame with HEVC encoder
13. Perform timing analysis
14. for m  1:16:h
15. for n1:16:w
16. Extract bits per frame block with mb header
17. for i m:4:m+15
18. for j n:4:n+15
19. if (i==1&j==1)
20. Perform No Optimization
21. else if (i==1)
22. Horizontal Prediction
23. else if (j==1)
24. Vertical Prediction
25. End
26. end
27. end
28. end
29. end
30. compute T1total,T2total,T3total
31. Append I-frame header with „1111‟
32. Extract features of I-frame ;
33. Encoding the video
34. Calculate the percentage of ET

101
 ET1 T1total/ (T1total+T2total+ T3total)
ET2 T1total/ (T1total+T2total+ T3total)
ET3 T1total/ (T1total+T2total+ T3total)
35. Store Rf
36. Append I-frame bits with „11100…‟
37. Construct reference frame for P-frames

End

The above algorithm III performs HEVC encoding integrated with H.265 where a
timing analysis (Line-13) computes the encoding time on consecutive frame sequence.
Apparently, the process also performs motion estimation (Line-4) as well as extracts the
features of the input video sequence (Line-5). The proposed encoding uses a level-2
optimization as well as different types of prediction mode on P-frame bit stream (Line-
1419). After successful completion of timing analysis the process append I-frame header
with „1111‟ (Line-32), then it also computes different mode of encoding time. Finally it
appends the receiving I-frames bit streams with „11100‟ (Line-36) and construct the
reference link for P-frames (Line-37).

6.5.1.3 Experimental Measurements for Algorithm-III

This segment represents the encoded images (i.e. Level-2 Optimization) obtained
after simulating the above algorithm in a parameterized test bed configuration. It shows
different frames (here 1-5) of a video sequence encoded with the proposed integrated HEVC
Level-2 video coding standard. The following are encoded video frames obtained for a
sample video file (302x223) of .avi extension.

Table 6.7 Numerical Simulation for Encoded Frames-Level-2

Frame Index Encoded Frames (Level-2)

1
(I-frame)

102
 2
(P-frame)

3
(P-frame)

4
(P-frame)

5
(P-frame)

The simulation for processing HEVC Level-2 optimization is explicitly carried out on
a real-time (not static) video frames before uploading the encoded bit stream to the network
layer. Further, the study also formulates another level of video frame encoding and
optimization scheme, to be precise the terminal layer of encoding pipeline enhance the
perceptual quality of the pixel entities while processing the compression very effectively.
The modelling of the proposed HEVC using H.265 pertaining to the fact that it also reduces
the computational complexities while processing video frames offline (for both encoding and
decoding scenario). However, the following algorithm-IV depicts the conceptual model of
the final encoding from a theoretical perspective and it also highlights its respective
simulation outcomes on the experimental measurement section. The proposed Algorithm-IV
basically continuance of the prior two algorithms for encoding the P-frames considering
block sizes of each frames, which is stated as follows.

103
 Table- 6.8Notation for Algorithm-IV

Sl. No. Symbolizations Description

1. bi Size of blocks
2. fl End frame
3. MSE Mean Square Error
4. µLM Lagrangian multiplier used for motion estimation
5. µLMp Lagrangian multiplier used for partition mode selection
6. mv Array containing previously computed motion vectors
7. s Array containing previous frame and the current frame
8. Sr The coded current frame
9. f The current frame to be coded
10. ext Flag to indicate whether to use extended cosine ME
11. is_ext Array for result visualization
12. MBsize Macro block size
13. Bsize

 Level-3 Encoding (Optimization)

Algorithm-IV: HEVC Encoding Pipeline with H.265 (Level-3)

Input: Vseq, Qp, bi

Output: EVframes (Level-3)

Start

1. Compute bi bi * 4;
2. for k2:fl
3. kk-1
4. Encode P-frame for H.265/HEVC
5. Append P-frame header with „0000‟
6. Extract P-frame
7. Global Qp, µLM, µLMp
8. global s , Sr, mv, f, ext, is_ext
9. find the height and width and number of frames in the Vseq (frame sequence)
10. Calculate µLM for the current value of Qp
11. Compute µLM
µLM0.85* (2^ ((Qp-12)/3))
12. µLMp(µLM).5
13. for f:2

104
 14. compute mv
15. for i1:bi: h
16. for j1:bi: w
17. Define bi
18. Define search range of mv
19. Determine the 16x16 partition mode
20. if (partition_mode< 4)
21. Write the reconstructed 16x16 macro block to the output array
22. Write block mv to the array
23. else
24. Find partition mode for 8x8 sub-block
25. Define macro block size
MBsizeBsize /2
26. Define motion vector search range
27. Add header for a quad split with „00100‟
28. Determine 8x8 partition mode for each sub block
29. Determine the 8x8 partition mode for each sub block
30. Write reconstructed sub block to the output array
31. Apply mode decision for different blocks
32. end
33. end
34. end
35. Reconstruct the frame seq.
36. end
37. Store receiving frame
38. Append P-frame bit stream with „1100…‟
39. Define reference frame for next P-frames

End

After the completion of the above process the proposed framework computes the
compression ratio by applying the following eq. 6.1.


CR = (6.1)


Where  denotes the number of original bits and  denotes the number of
compressed bits. Finally it saves the encoded video stream into a form of matrix (stream of
bits). The further process includes initializing DASH server and transmission of the encoded
video sequence using DASH server. The algorithm-IV performs H.265 encoding on the P-
frames by appending it with header info „0000‟ (Line-5). Further it extracts the features P-
frame (Line-6) followed by followed by computing the motion estimation using Lagrangian

105
multiplier as well as partition block selection (Line- 11,12). It also accesses different
partition mode for determining the block sizes and further reconstructs the block sizes within
a search range of motion vectors. Finally it reconstructs the frame sequence along with
respective sub-block formation and receiving frames are stored after being appended with
„1100…„ (Line-38) .The following are the experimental measurements evaluated by
processing the Algorithm-IV in a parameterized simulation environment.

6.5.1.4 Experimental Measurements for Algorithm-III

Despite of performing the Level-3 encoding for each frame of an .avi extension, the
proposed algorithm-IV achieves significant high end compression by maintain quality
information. The following are the encoded images from frame 1 to 5 of .avi video object.

Table 6.9 Numerical Simulation for Encoded Frames-Level-2

Frame Index Encoded Frames (Level-3) Each Frame Size

1
302x223
(P-frame)

2
(reference 302x223
P-frame)

3
302x223
(P-frame)

4
302x223
(P-frame)

106
 5
302x223
(P-frame)

6.5.1.5 Initiating DASH Server (TCP/IP Protocol Stack) for Video Transmission

This segment introduces the design aspects of the DASH server and its significant
contribution on transmitting multimedia files efficiently. It has gained the attention of
researchers for integration of its features in a larger scope. The proposed study also adopts
the design methodology of DASH server for transmitting real-time encoded videos using
HTTP based TCP/IP protocol stack. However, the prime characteristics of the DASH server
lie into the fact that it can be integrated with any network protocols irrespective of any
configuration issues. The extensibility of this adaptive bit rate streaming technology helps in
transmitting the real time streaming videos by partitioning it into one or more segments. In
the recent scenario most of the conventional online video streaming and compression
protocols such as H.265, H.264 etc. started adopting HTTP based DASH servers for the ease
of multimedia file transmission along with the scope of extensibility to a larger extent. The
proposed system intended to transmit the encoded video frames over an wireless channel
without utilizing more band-width thus it incorporates streaming of encoded 8K (UHD)
video frames in real-time for on-demand services. However, the key idea behind the
proposed concept for video encoding and transmission of the encoded video over wireless
medium has been initiated by referring conventional design flaws associated with HTTP/1.1
web server based uploading and downloading methods. Few design considerations for HTTP
based DASH are further stated in this context, are as follows.

a. The proposed framework implemented over a wireless mobile communication link

and thereby, it should also ensure the reliability and ease of communication protocol
deployment (i.e. TCP/IP Protocol Deployment) for an efficient connection
establishment in between client and server units.
b. The proposed HEVC enables real-time video streaming on mobile devices by
installing the features of H.265 standard which is compatible with the 8K UHD video
processing units.

107
 This led to a scenario of adopting H.265 in mobile communication and video
processing standards. The proposed framework also incorporates a special type of network
namely Content Distribution Network (CDN) for effectively transferring the compressed
multimedia files without affecting the bandwidth constraints and quality of the data. It also
minimizes the downloading latency while processing the adaptive bit rate switching for real-
time video streaming services.

However, once the DASH server gets initiated in the server unit, it attempts to locate
the IP and port number of the client devices trying to establish communication with the
server unit for having access to the encoded video sequence. It waits for the clients to
respond for specific time duration. Once the encoding process is finished for the video
sequence, then the video object get transferred to the DASH server. Further, the respond
from client side to the server side establishes the communication using Host IP, and Port
number, e.g. the proposed system uses TCP/IP protocol stack as a communication function
subjected to enable support for server socket. It also takes the remote host IP and port
number as functional arguments. Meanwhile, after the successful connectivity establishment
in between client and server unit, client start receiving the encoded video sequence and
perform HEVC decoding simultaneously to retrieve and view the original video. The
experimental analysis of the overall sequence ensures the fact that our proposed framework
achieves a very effective compression ratio while maintaining the video quality
simultaneously.

The following algorithm showcases the step by step procedure to upload the encoded
bit stream into DASH server, once it gets initiated.

Table- 6.10 Notation for Algorithm-VI

Sl. No. Symbolizations Description

1. bi Size of blocks
2. Vseq Video sequence
3. t Binary file storing time duration
4. Seg Segment value

108
Algorithm-V: Algorithm for uploading entire HEVC encoded bit stream into DASH

Input: EVframes

Output: Entire bit stream Vsequploaded

Start

1. Global t
2. Load bitstream
3. Define Seg
4. Compute chunk 20*seg
5. While 1
6. Compute length(bits)
7. if i+chunk>length(bits)
8. data = bits(i:length(bits)
9. Write data into binary file
10. break;
11. end
12. Compute data from the entire bitstream
13. ii + chunk
14. write the entire bitstream into output block
15. end
16. Entire bit stream uploaded

End

The algorithm-V clearly depicts the scenario of uploading HEVC encoded bitstream
into the DASH server. The above mentioned algorithm processes an encoded bitstream after
dividing it into one or more segment (Line-3). Further it computes the length of the stream
and compares it with a standard buffer value. If the process finds the amount of bitstream is
lesser than the standard buffer value thus it writes the data into a binary file (Line-69).
Finally it uploads the entire bit stream into the DASH server. The next section performs a
comparative analysis based on the outcomes being evaluated after simulating the framework
in a test bed.

6.6 COMPARATIVE ANALYSIS

As the proposed HEVC framework has been integrated with the most recent H.265
protocol to enhance the perceptual quality of a compressed video file, therefore the use of
multi-level optimization along with the video coding makes it more efficient and less

109
complex. Moreover, the design and development of the proposed system has been carried out
considering two different instances of simulation tool where one instance played the role of
server unit, performs the HEVC Encoding and install the DASH server respectively to
transmit the encoded video file. On the other hand, another instance acted like a client
module, performed decoding on the received video sequence. During the comparative
analysis our proposed study considered number of processed frames and PSNR (dB) as
performance parameters. It basically finds the perceptual quality of the 8K (UHD) video
frames after processing through a high end compression. The results obtained after
simulating proposed H.265 based HEVC protocol shows that it produces the best outcomes
so far and able to exhibits the conventional H.264 and H.263 in terms of signal quality
(PSNR) and computation time. The following Figure 6.2 shows a comparative study to
measure the perceptual quality of different Compression techniques.

Table 6.11 Numerical Outcomes for

Perceptual Quality Evaluations
Number of 50
Processed 40
PSNR (dB)

Frames (n) H.263 H.264 H.265 30

1 23.12 27.46 30.2 20 H.263
2 20.24 27.46 32.75 10 H.264
3 25.32 30.25 38.34 0
Proposed
4 20.12 32.78 38.67 1 2 3 4 5 6
5 20.12 28.64 42.2 Number of Processed Frames (n)
6 25.87 35.1 42.2

Figure 6.2 Perceptual Quality Evaluations

The Figure 6.2 clearly shows the effectiveness of the proposed framework on
achieving high-end compression along with maintaining superior quality of encoded video
frames i.e. 8K resolution of UHD video frames. It shows that the proposed model achieves a
PSNR of 35 to 42 dB after processing 1-6 frames while the curve for H.264 tends to remain
fix in the upper bound of 35 dB. More or less the H.263 also achieves very less amount of
superiority (20 dB- 25dB) in the pixel components due to lack of optimization. The proposed
study also performs another comparative analysis to evaluate the encoding time, taken by
different compression techniques and our proposed model. It mostly compares the outcomes
(i.e. Processing Time) obtained after encoding a 8K UHD video sequence using H.263,
H.264 and the proposed framework.

110
Table 6.12 Numerical Outcomes for 0.1

Processing Time (sec)

0.08
Processing Time vs No of Frames
0.06
No of 0.04 H.263
Frames (n) H.263 H.264 Proposed 0.02 H.264
1 0.018 0.024 0.016 0
Proposed
2 0.02 0.029 0.0165 1 2 3 4 5
3 0.024 0.03 0.0175 No of Frames (n)
4 0.057 0.033 0.018
5 0.083 0.038 0.019
Figure 6.3 Processing Time vs No of
Frames

Figure 6.3 also shows that the proposed framework achieves better outcomes by
taking very less amount of processing time (to be precise encoding time) in comparison with
the current baselines such as H.263 and H.264. It leads to having a significant impact on
enhancing the performance of the advanced H.265 based HEVC protocol. The significant
outcomes of the proposed study ensure its effectiveness on real-time video compression,
irrespective of losing quality of data.

6.7 CONCLUSION

This chapter is fundamentally formulated to conceptualize a novel video coding

framework based on HEVC encoding. However, the novelty of the proposed concept is, it
implies three different levels of optimization methodologies to enhance the perceptual quality
of UHD video frames while accessing the encoding pipeline. The outcomes of the study
show that it significantly performs intra and inter prediction in between blocks of a frame to
determine high resolution pixel entities. The comparative analysis performed in this chapter
measures its effectiveness by comparing its performance metric with conventional baselines
such as H.264/H.263. Finally, the study concludes that the proposed framework outperforms
the conventional techniques and achieves very high compression ratio along with superior
signal quality (i.e. PSNR range in between (30 dB - 40 dB). The study also depicts that our
proposed model reduces the computational complexities during encoding phase as compared
to the state of art techniques.

111