Abstract—Ringing effect is one of the most annoying visual artifacts in digital video. It is a significant factor of subjective quality deterioration. However, there is a widely-accepted misunderstanding of its cause. In this paper, we propose a reasonable interpretation of the cause of ringing effect. Based on the interpretation, we suggest further two methods to reduce ringing effect in DCT-based video coding. The methods adaptively adjust quantizers according to video features. Our experiments proved that the methods could efficiently improve subjective quality with acceptable additional computing costs. Keywords—ringing effect, video coding, subjective quality, DCT. I. INTRODUCTION ue to the great advancements of digital video compression techniques in the past ten years, digital video has already melted into our everyday lives omni-directionally. However, digital video has some exclusive artifacts never seen in analog video. Ringing effect is such a notorious artifact, which is often complained by videophiles but somehow overlooked by many experts and researchers. Ringing effect means haloes and/or rings near sharp object edges in the picture. It is very well known in DVD fans. In fact, it is mentioned in almost every DVD review [1] recently. In many DVD movies, including both old movies and new movies, you can notice ringing effect every now and then. Fig. 1 shows an example picture picked from the “Thelma & Louise: Special Edition” DVD. Between the chine and the sky, there is a weird outer ring along the chine, which is so-called ringing effect. Ringing effect will become more obvious and annoying when viewed on big screen display devices, such as projectors. Experts usually prefer objective quality to subjective quality for measurement. The measurement of objective quality needs “original” video [2]. However, the ultimate purpose of video is for human viewing. In most cases, audience cannot compare what they see with “original” video, since the “original” video is not available. They can only compare what they see with pictures in their experiences and memories. In these cases, Manuscript received November 4, 2003. This work was sponsored by the Australian Research Council Grant and the Chinese National Natural Science Grant 60273008. Y. Yuan is with the Department of Computer Science, Tsinghua University, Beijing 100084, P. R. China. (phone: +86 10 62771545; e-mail: yyu99@mails.tsinghua.edu.cn). D. Feng is with the School of Information Technology, University of Sydney, NSW 2006, Australia. (e-mail: feng@staff.cs.usyd.edu.au). Y. Z. Zhong is with the Department of Computer Science, Tsinghua University, Beijing 100084, P. R. China. (e-mail: zyz-dcs@tsinghua.edu.cn). subjective quality is as important as objective quality. If we achieve higher subjective quality with similar objective quality, both experts and mass audience will be satisfied. If we use PSNR as the measure of objective quality, we will not find noticeable objective quality deterioration caused by ringing effect, since the haloes and rings only occupy very small areas. However, no one has ever seen such weird haloes and rings in real life, so ringing effect will be very noticeable in the pictures, i.e., will cause significant subjective quality deterioration. It is necessary to find a way to reduce ringing effect. The rests of this paper are organized as follows. In section 2, we analyze the cause of ringing effect in detail. In section 3, we present two methods for ring effect reduction in DCT-based video coding. In section 4, we describe our experiments and results. Finally, in section 5, we discuss the usability of our method, and forecast its application foreground. II. THE CAUSE OF RINGING EFFECT In order to find a solution to address ringing effect, we should The Causation and Solution of Ringing Effect in DCT-based Video Coding Yu Yuan, David Feng, and Yu-Zhuo Zhong D FIGURE I. RINGING EFFECT EXAMPLE FROM DVD MOVIE World Academy of Science, Engineering and Technology International Journal of Computer and Information Engineering Vol:2, No:1, 2008 231 International Scholarly and Scientific Research & Innovation 2(1) 2008 ISNI:0000000091950263 Open Science Index, Computer and Information Engineering Vol:2, No:1, 2008 publications.waset.org/3309/pdf