Printer Friendly

H.264/AVC DEBLOCKING FILTER FOR 1080P HIGH DEFINITION VIDEO.

Byline: G. Raja - Email: gulistan.raja@uettaxila.edu.pk

ABSTRACT: High Definition Video (HDV) is becoming an emerging application due to desire of high display resolution and superior video quality. Video compression is used for efficient transmission of HDV over network. However, loss of correlation between adjacent blocks during quantization process in compression results in visually disturbing artifacts called blocking artifacts. The latest H.264/AVC incorporates deblocking filter for reduction of blocking artifacts. This paper describes performance comparison of H.264/AVC deblocking filter for 1080p high definition video. Various high definition video sequences of typical content have been used for experimentation. The results show that H.264/AVC deblocking filter successfully suppresses the blocking artifacts in HDV without significantly degrading the objective quality of video.

Key words: high definition video, blocking artifacts, deblocking filter, H.264/AVC, 1080p

INTRODUCTION

The need of video communications is increasing day by day with the exponential growth of Internet. As a result, high definition (HD) digital video and video over wireless networks are becoming emerging applications due to higher quality (Alvarez et al., 2005). Some of high definition applications are in Blue-ray, HD-DVD, HDTV, IPTV, digital cable, dish HD, Intenet HD downloads etc. High definition standard is developed by ATSC (Advanced Television Systems Committee) and adopted by FCC (Federal Communications Commission) in 1996 (Wu et al., 2006).

There are two commonly used display resolutions for HD video: (1) 1080p (1920 x 1080 pixels) and (2) 720p (1280 x 720 pixels). In 1080p display resolution, 1080 stands for number of horizontal scan lines and 'p' stands for progressive scan (complete frame is scanned sequentially from top to bottom in horizontal direction) whereas in 720p resolution, there are 720 horizontal scan lines of display.

The bandwidth required to transmit raw HD video without using video compression is around 4 Gbps (Shimu, 2010). Therefore, video compression is necessary for transmission of HD video over wireless networks. These applications not only require higher compression efficiency for efficient transmission over network but also higher video quality is required.

H.264/AVC is the latest international video compression standard designed for variety of applications from low bit rate internet streaming to HDTV broadcast and digital cinema applications. Performance analysis of latest H.264/AVC standard showed its significant coding gains and compression efficiency over existing video compression standards (Weigand et al., 2003; Raja and Mirza, 2004) However, during compression process, loss of correlation between adjacent blocks during quantization process results in visually disturbing artifacts called blocking artifacts.

Motion compensated prediction is another source of blocking artifacts in video. nterpolated pixel data from different locations in reference frames are used for generating motion compensated blocks. As a result, there is almost never a exact match for this data and this introduces discontinuities on edges of copied blocks. Moreover, visual discontinuities also appear during copying process where existing edge discontinuities in reference frames are carried into interior of blocks to be compensated. The H.264/AVC standard employs deblocking filter for reduction of blocking artifacts while contributing to bit rate saving. Research reported in literature shows that H.264/AVC deblocking filter can successfully suppress blocking artifacts at low bit rates (List et al., 2003; Richardson, 2003) This paper describes in-depth performance evaluation of H.264/AVC deblocking filter for 1080p high definition video for IPTV and video over wireless.

Overview of H.264/AVC Deblocking Filter: H.264/AVC video compression algorithm employs in- loop deblocking filter after the inverse transform in the encoder and decoder (Puri et al, 2004). Vertical and horizontal edges of block in each macroblock (MB) of a frame are filtered for suppression of blocking artifacts. Luminance and chrominance components of video are separately filtered by H.264/AVC deblocking filter. The filtering operation affects three samples on both sides of edges on a block boundary (List et al., 2003; Richardson, 2003; Raja and Mirza, 2006). The four samples, p0, p1, p2, p3 and q0, q1, q2, q3 on horizontal and vertical edge in adjacent blocks of a macroblock are shown in Fig-1.

The operation of deblocking filter can be divided into two main steps, i.e., boundary (filter) strength calculation and filter application respectively.

The decision to apply filter on samples of block edges is taken with the help of boundary strength (bS) parameter. The bS is computed for each edge between adjacent 4 x 4 luminance blocks and assigned an integer value from 0 to 4.

Same value of boundary strength computed for luminance blocks is used for chrominance blocks. Various parameters are used for computation of boundary strength which includes macroblock type, motion vector, quantization parameter, gradient of samples across edges (Richardson, 2003; Raja and Mirza, 2006).

The decision to apply filter on samples of block edges does not only depend on non-zero boundary strength. Application of deblocking filter may not be needed even in the case of non-zero boundary strength. For example, filtering is not required in case of real edges in the video as application of filter will do blurring and it may result in difficulty in distinguishing real edges. On the other hand, samples do not have much variation in smooth regions of video and appea nce of artifacts is more obvious. Therefore, an additional condition other than non-zero boundary strength is needed for efficient application of deblocking filter. Consequently, block edge samples (p2, p1, p0, q0, q1, q2) are filtered only if they meet the following conditions (ITU-T, 2003).

RESULTS AND DISCUSSION

Evaluation of H.264/AVC deblocking filter for suppression of blocking artifacts is carried out using joint reference model software version 10.2 (Karsten, 2010). The subjective and objective quality without- and with deblocking filter is measured.

Various 1080p high definition video test sequences of different scenarios are used for analysis. The standard sequences used for experimentation are BLUE SKY, DUCKS TAKE OFF, PEDESTRIAN AREA, RIVER BED, RUSH HOUR, STATION, SUNFLOWER, TRACTOR (HDV 1080p, 2010). Each sequence is coded at five different bit rates consists of 50 frames. The other various parameters incorporated in simulation model are:- frame rate: 30, GOP structure: IPPP, B frames not used, search range: 16, no. of reference frames: 5, entropy coding method: CABAC. The coding performance is measured by comparing PSNR (peak signal to noise ratio) with- and without deblocking filter. Table-1 shows the objective comparison of deblocking filter by computing luminance PSNR at different bit rate for various 1080p sequences. Objective results indicate slight improvement in PSNR by use of deblocking filter in comparison without using deblocking filter.

Table-i. Average luminance PSNR at different bitrates for iO8Op sequences with- and without deblocking filter

1080p###Bit-Rate###YPSNR (dB)###Bit-Rate###YPSNR (dB)

Sequence###(Mbps)###No###With###1080p###(Mbps)###No###With Filter

###Filter###Filter###Sequence###Filter

###5.7###22.19###22.20###4.2###23.56###23.72

###6.8###22.68###22.71###4.9###24.43###24.57

Ducks

###8.7###23.25###23.25###Sunflower 6.1###25.27###25.35

Take Off

###11.9###23.91###23.93###7.7###26.14###26.20

###17.2###24.71###24.73###10.1###27.06###27.09

###4.3###22.93###23.01###5.5###22.83###22.93

###5.0###23.54###23.64###6.7###23.41###23.46

Pedestrian

###6.2###24.27###24.35###Tractor###8.6###24.02###24.07

Area

###7.9###25.08###25.16###11.4###24.76###24.84

###10.5###26.01###26.06###15.4###25.64###25.69

###3.9###24.86###24.95###2.4###25.95###26.06

###4.4###25.53###25.58###3.2###26.49###26.55

Rush Hour###5.3###26.24###26.31###Station###4.4###27.02###27.21

###6.7###27.01###27.06###6.4###27.72###27.81

###8.8###27.88###27.88###9.4###28.46###28.62

###3.5###22.71###22.75###6.8###22.37###22.47

###4.4###23.29###23.35###7.9###23.04###23.14

RiverBed###5.8###23.91###23.98###BlueSky###9.8###23.77###23.82

###7.9###24.69###24.72###12.5###24.57###24.58 (a)

###11.1###25.53###25.55###16.3###25.45###25.44

Objective comparison of luminance component window of DUCKS TAKEOFF frame 4 encoded at 5.7 Mbps without and with H.264/AVC deblocking filter is shown in Fig-3. It has been observed that H.264/AVC deblocking filter can do substantial suppression of blocking artifacts. However, granular like noise appears and appearance of this noise overrides the effectiveness of deblocking filter in high definition video sequences at lower bit rates.

Conclusion: Performance evaluation of H.264/AVC deblocking filter is presented for 1080p high definition video. Different test video sequences are used for analysis. Simulation results show that H.264/AVC deblocking filter can significantly reduce the visually disturbing blocking artifacts. However, appearance granular like noise reduces the efficiency of deblocking filter for 1080p high definition video.

REFERENCES

Alvarez, M., E. Salami, A. Ramirez and M. Valero. A performance characterization of high definition digital video decoding using H.264/AVC. Proc. of the IEEE Intl. Workload Characterization Symposium, USA: 24 - 33 (2005).

HDV 1080p, High Definition Video Test Sequences. Institute for Data Processing. Department of Electrical Engineering and Information Technology, Technical University Munich, Germany (2010).

ITU-T, Joint Video Team. Draft ITU-T recommendation and final draft international standard of joint video specification, ITU-T Rec. H.264 | ISO/IEC 14496-10 AVC. ISO/IEC MPEG and ITU-T VCEG, JVT-G050r1 (2003).

Karsten, S. H.264/AVC Software Coordination, Joint Model Software Version 10.2. Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, Germany (2010).

List, P., A. Joch, J. Lainema, G. Bjontegaard and M. Karczewicz. Adaptive deblocking filter. IEEE Trans. on Circuits and Systems for Video Technology, 13: 614-619 (2003).

Puri, A., X. Chen and A. Luthra. Video Coding using the H.264/MPEG-4 AVC Compression Standard. Signal Processing: Image Communication, 19:793-849 (2004).

Raja, G., and M. J. Mirza. Performance comparison of advanced video coding H.264 standard with baseline H.263 and H.263+ standards. Proc. of 4th IEEE Intl. Symposium on Communications and Information Technologies, Sapporo, Japan: 743-746 (2004).

Raja, G., and M. J. Mirza. In-loop deblocking filter for JVT H.264/AVC. WSEAS Trans. on Signal Processing, 2:143-148 (2006). Richardson, I. E. G. H.264 and MPEG-4 Video Compression. Wiley and Sons Ltd, England (2003).

Shimu, S. S. Performance analysis of H.264 encoder for high-definition video transmission over ultra- wideband communication link. MS Thesis, Univ. of Saskatchewan, Saskatoon (2010). Weigand, T., G. J. Sullivan, G. Bjontegaard and A. Luthra. Overview of H.264/AVC video coding standard. IEEE Trans. on Circuits and Systems for Video Technology, 13:560-576 (2003).

Wu, Y., S. Hirakawa, U. H. Reimes and J. Whitaker. Overview of digital television development worldwide. Proc. of the IEEE, 94:8-21 (2006).

Electrical Engineering Department, University of Engineering and Technology, Taxila
COPYRIGHT 2011 Asianet-Pakistan
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Pakistan Journal of Science
Article Type:Report
Geographic Code:9PAKI
Date:Mar 31, 2011
Words:1829
Previous Article:IMPACT OF BIRD FLU ON DEMAND OF CHICKEN MEAT IN FAISALABAD.
Next Article:ON BAYESIAN ANALYSIS OF THE EXPONENTIAL SURVIVAL TIME ASSUMING THE EXPONENTIAL CENSOR TIME.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters