By Sudeep Gangavati ID 1000717165 EE5359 Spring 2012 UT Arlington Objective and motivation The goal to compare H264AVC AVS P2 and Dirac Video quality assessment MSE PSNR SSIM ID: 731009
Download Presentation The PPT/PDF document "Video Compression Standards : A Comparat..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Video Compression Standards : A Comparative Analysis of H.264, Dirac and AVS P2
By
Sudeep
Gangavati
ID 1000717165
EE5359 Spring 2012,
UT ArlingtonSlide2
Objective and motivation
The goal – to compare H.264/AVC, AVS P2 and Dirac
Video quality assessment – MSE, PSNR, SSIM
Ever
increasing demand for video
compression
Several
different video coding standards have been developed to address the needs efficient video coding for multitude of applications like video streaming, TV broadcasting, 3D TV, Free-viewpoint TV etc.
[24]. Slide3
H.264/AVC Features
The most widely used video coding standard
Fig. 1 Video coding standards evolution [12]Slide4
FeaturesMotion compensated coding structure
Picture
slices MBs
subMBs
blocks
pixels. This is shown in Figure 3.
Only 4:2:0
chroma
format was
supported
earlier and 4:2:2 ,
4:4:4 were added
later. This is shown in Figure 2.
I , P and B slices
Derived slices SI and SP Slide5
Fig 2.
4:4:4,
4:2:2
,
4:2:0
sampling patternsSlide6
Fig 3. H.264 syntaxSlide7
Profiles and levels
Main Profile
Baseline Profile
Extended Profile
High ProfileSlide8
H.264 Profiles
Fig.4 H.264 profiles [1]Slide9
H.264 Encoder
Fig. 5 Encoder structure for H.264 [2]Slide10
H.264 Decoder
Fig.6 Decoder structure of H.264 [2]Slide11
Intra and Inter Predictions
Intra Prediction :
Uses
spatial prediction to reduce spatial redundancy.
4 X 4
luma – 9 modes 16 X 16
luma
– 4 modes
8
X
8
luma
-
9
modesSlide12
Intra prediction modes for 4X4
luma
Fig.7(a) Intra prediction modes [6]
The samples above and to the left,
labelled
A-M in Figure 7 have previously been encoded and reconstructed and are therefore available in the encoder and decoder to form a prediction reference.Slide13
Intra Prediction Modes for 16x16 luma
Again the previously encoded samples directly above and to the left of the
macroblock
have been reconstructed and are used for the prediction
Fig 7 (b) Intra prediction modes for 16x16
luma
[6]Slide14
Inter prediction
Uses motion estimation and motion compensation (MC).
Fig.8 H.264 Inter prediction [5]Slide15
De-blocking filter[5]
Is used to reduce the blocking artifacts.
Since
the filter is present in the loop , it prevents the propagation of the blocking artifacts
.
Fig. 9
Boundaries in a
macroblock
to be filtered (
luma
boundaries shown with solid lines
and
chroma
boundaries
shown with dotted
lines) [1]Slide16
AVS China[7]
AVS-Audio Video Standard
Standardization includes system, audio, video and digital copyright management.
Goal – to achieve coding efficiency with reduced complexity. Slide17
AVS Parts [3]
Fig. 10 AVS China parts [3]Slide18
AVS P2 Encoder [7]
Fig. 11 AVS part 2 encoder [7]Slide19
AVS P2 decoder
Fig 11 (a) AVS P2 decoder block diagram [7]Slide20
Intra Prediction in AVS[7]
Spatial prediction is used in intra coding in AVS part 2.
The Intra prediction is based on 8x8 block
The intra prediction method is derived from the neighboring pixels in left and top blocksSlide21
Intra Prediction contd.
Fig.12 (a) Five different modes for intra luminance prediction[16]Slide22
Inter prediction [16]
Inter prediction in AVS is by motion compensation and motion estimation [16].
As shown in the Figure 12 (b), the
macroblock
can have 16 x 16
, 8 x 16, 16 x 8 or 8 x 8 [16].
Fig 12 (b)
Macroblock
sizes [16]Slide23
Dirac
Dirac is a video codec originally developed by BBC
This technique is used from web streaming of videos to HD TV applications to storage of content.
Dirac can compress any resolution picture
The encoder and decoder diagrams are shown in Figure 13 (a) and (b) respectively.Slide24
Figure 13 (a) Dirac encoder[8]
Figure 13 (b) Dirac decoder[8]
Dirac encoder and decoder : Slide25
Dirac pro Features
Dirac
pro
supports the following technical aspects
[9]:
Intra-frame coding only 10 bit 4:2:2
No s
ubsampling
Lossless or visually lossless compression
Low latency on encode/decode
Support
for multiple HD image formats and frame rates
Low complexity for decodingSlide26
Experimental Results
Implementation of DIRAC Software 1.02:
Video sequence:
news_qcif.yuv
.
Width: 176, Height: 144.
Total number of frames: 300
Number of frames used for encoding: 100.
Frame rate: 25 FPS. File Size: 3713kB
.
Table
1: Parametric values for Dirac video
codec
Quality Factor
Compressed File Size
Bit rate (kBps)
Y-PSNR(dB)
Y-MSE
Y-SSIM
Comrpession Ratio
0
38
9.573
25.773
187.13
0.79
98:1
5
61
15.571
32.134
39.588
0.95
60:1
10
369
96.301
46.699
1.41
0.98
10:1
15
1278
130.12
51.799
0.743
0.99
2.9:1Slide27
Experimental Results contd.
Fig.14 Output
of Dirac video codec at
different Quality Factors
Quality Factor = 0
Quality Factor = 5
Quality Factor = 10Slide28
Video
sequence:
foreman_qcif.yuv
Width:
176;
Height: 144.
Total number of frames: 300
Number of frames used for encoding: 100.
Frame rate: 25
FPS; File
Size:
3713kB
Quality Factor (QF)
Compressed File Size (kB)
Bitrate (kBps)
Y-PSNR (dB)
Y-MSE
Y-SSIM
Compression Ratio
0
27
8.991
21.5
301.56
0.6875
138:1
5
58
12.675
28.91
110.12
0.8613
64:1
10
581
140.673
43.675
0.827
0.979
6:1
15
1340
170.342
49.556
0.667
0.99
2.7:1
Table 2. Parametric values of Dirac video codec for
foreman_qcif
videoSlide29
QF=0
QF=05
QF=10
QF=15
Fig.15
Output of Dirac video codec at different Quality
Factors for
foreman_qcif
video Slide30
Implementation of AVS softwareVideo sequence used:
news_qcif.yuv
Width: 176;Height: 144
Total number of frames: 300; Number of frames used: 100
Frame rate: 25 FPS; File Size: 3713kB
QP=0 QP=10 QP=50
Fig.16
Output of
AVS video
codec at different Quality
Factors for
news_qcif
video
Fig.17
Output of
AVS video
codec at different Quality
Factors for foreman _
qcif
videoSlide31
Implementation of AVS software
Quantization Parameter (QP)
Compressed File Size
Bit rate (kBps)
Y-PSNR(dB)
Y-MSE
Y-SSIM
Comrpession Ratio
0
980
554.19
54.773
0.2587
0.9997
3:1
10
442
219.12
49.72
0.5525
0.9945
9:1
30
64.0
156.49
38.49
9.23
0.9760
58:1
50
12.0
29.26
27.96
104.13
0.8506
309.7
Table 3. Parametric values of AVS video codec for
news_qcif
video
QP
Compressed file size
Bitrate
Y-PSNR
Y-MSE
Y-SSIM
Compression Ratio
0
1123
478.88
52.658
0.2823
0.998
3:1
10
450
278.9
48.775
0.781
0.9903
8.25:1
30
70
79.66
35.231
13.56
0.867
53:1
50
14
24.5
29.780
146.32
0.776
265:1
Table 4. Parametric values of AVS video codec for
foreman_qcif
videoSlide32
Implementation of H.264 software (JM 18.0)
Table 5. Parametric values of H.264 video codec for
news_qcif
video
Table 6. Parametric values of H.264 video codec for
foreman_qcif
video
Quantization Parameter (QP)
Compressed File Size
Bit rate (kBps)
Y-PSNR(dB)
Y-MSE
Y-SSIM
Comrpession Ratio
0
279
685.19
60.773
0.21619
0.999
13:1
10
208
410.21
48.545
0.9653
0.9947
17.8:1
30
123
155.62
35.721
17.4211
0.8626
30:1
50
35
29.49
28.736
224.23
0.7644
27.5:1
Quantization Parameter (QP)
Compressed File Size
Bit rate (
kBps
)
Y-PSNR(dB)
Y-MSE
Y-SSIM
Comrpession Ratio
0
379
485.19
62.773
0.21619
0.999
9:1
10
210
340.21
54.67
0.7769
0.9947
18:1
30
98
155.62
34.721
14.4211
0.8626
39:1
50
25
39.49
27.736
236.23
0.6944
27.5:1Slide33
Plots of PSNR (dB) vs. Bitrate (kBps)
Fig18.
Plot of PSNR
vs.
Bitrate for different codecs for
news_qcif
videoSlide34
Plots of PSNR (dB) vs. Bitrate (kBps
) contd..
Fig 19.
Plot of PSNR
vs.
Bitrate for different codecs for
foreman_qcif
videoSlide35
Plots of MSE vs. Bitrate
Fig 20.
Plot of MSE
vs.
Bitrate for different codecs for
news_qcif
videoSlide36
Plots of MSE vs. Bitrate contd..
Fig 21.
Plot of MSE
vs.
Bitrate for different codecs for
foreman_qcif
videoSlide37
Plot of SSIM vs. Bitrate
Fig 22.
Plot of
SSIM vs.
Bitrate for different codecs for
news_qcif
videoSlide38
Plot of SSIM vs. Bitrate
Fig 23.
Plot of
SSIM vs.
Bitrate for different codecs for
foreman_qcif
videoSlide39
Computational complexity
Fig 24.
Plot of time taken by each codec at QP=30 and QF=10Slide40
Conclusions
The plots and tabulations show that with the increase in bitrate, there is an increase in PSNR and SSIM and reduction in the MSE.
Therefor from the plots and the tables, it can be concluded that H.264 provides optimum performance with respect to PSNR, MSE and SSIM over AVS part 2 and Dirac.
Regarding the computational complexity, H.264 is more complex than the other two standards viz., AVS part 2 and Dirac. This is due to the fact that H.264 supports several prediction modes and has varied
macroblock
sizes when compared to AVS and Dirac.Slide41
References
[1] Soon-
kak
Kwon, A.
Tamhankar
and K.R.
Rao
, “Overview of H.264 / MPEG-4 Part 10 (pp.186-216)”, Special issue on “ Emerging H.264/AVC video coding standard”, J. Visual Communication and Image Representation, vol. 17,
pp.186-216,
April 2006
.
[2]
T.
Wiegand
, G. Sullivan, G.
Bjontegaard
and A.
Luthra
, “Overview of the
H.264/AVC video coding standard
,” IEEE Trans. on Circuits and Systems for Video Technology
,
vol. 13, pp.560-576, July 2003
.
[3]
T.
Sikora
, “Digital video coding standards and their role in
video communications”, Signal Processing for Multimedia, J.S. Byrnes (Ed.), IOS press, pp. 225-251, 1999.
[4]
K. R.
Rao
, and D. N. Kim, “Current
video coding standards
: H.264/AVC, Dirac, AVS China and VC-1,” IEEE 42nd Southeastern symposium on system theory (SSST), March 7-9 2010, pp. 1-8, March 2010.[5]Z. Wang and A.C. Bovik
, “A universal image quality index”, IEEE Signal Processing Letters,Vol.9, pp. 81-84, March 2002.
[6] Iain Richardson, “ The H.264 advanced video coding standard”, Second
Edition,Wiley
, 2010
[7]
L. Yu et
al,
“An Overview of AVS-Video: tools, performance and complexity”, Visual Communications and Image
Processing,
Proc. of SPIE, vol. 5960, pp
. 679-690,
July
2005.
[8]
“ The Dirac web page” :http://
www.bbc.co.uk/rd/projects/dirac/intro.shtml
[9]
“Dirac Codec Wiki Page ” at http://en.wikipedia.org/wiki/Dirac(codec
)
[10]“Dirac
Pro web page” at http://
www.bbc.co.uk/rd/projects/dirac/diracpro.shtml
[11]
“Video on the web “
a http://etill.net/projects/dirac_theora_evaluation/
[12]
J.Lou
“
Advanced video codec optimization
techniques”, Doctoral Dissertation, Electrical Engineering Department, University of Washington, August 2009Slide42
References
[
13]
H.264 AVC JM Software : http://iphome.hhi.de/suehring/tml
/
[14]
H.264 decoder: http://www.adalta.it/Pages/407/266881_266881.jpg
[15]
W.
Gao
et al
,
“AVS - The Chinese
next-generation video coding
Standard” NAB, Las
Vegas, 2004.
[16]
X. Wang
et al
., “Performance comparison of AVS and H.264/AVC video coding standards” J.
Comput
. Sci. & Technol., vol.21, No.3, pp.310-314, May 2006.
[17]
AVS China part 2 video software, password protected : ftp://124.207.250.92/ [18] S. Swaminathan and K.R. Rao, “Multiplexing and demultiplexing of AVS CHINA video with AAC audio,” TELSIKS 2011, Nis, Serbia, 5-8 Oct. 2011. [19] Dirac Pro Software : http://diracvideo.org/download/
[20] M. Tun, K.K. Loo and J. Cosmas, “Semi-hierarchical motion estimation for the Dirac video codec,” 2008 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, pp.1–6, March 31-April 2, 2008. [21] T. Davies, “The Dirac Algorithm”: http://dirac.sourceforge.net/documentation/algorithm/,
2008. [22] Dirac video codec – A programmer's guide: http://dirac.sourceforge.net/documentation/code/programmers_guide/toc.htm [23] A. Ravi and K.R. Rao, “Performance analysis and comparison of the Dirac video codec with H.264 / MPEG-4 Part 10 AVC,”IJWMIP, vol.4, pp.635-654, No.4, 2011.[24] Proceedings of the IEEE Special issue on Frontiers of Audiovisual Communications, vol. 100, No.4, April 2012