Article Source: http://www.h3c.com.cn/products__technology/products/ip_multimedia/ip_video/HOME/Video_information/video_base/200802/334145_30003_0.htm
- Coding protocol and bandwidth selection of video conference
Summary: What kind of coding protocol should be used in the video conference? What is the main difference between the different H. Series Agreement? What kind of connection is there in different agreements and bandwidth? These are all current problems that often plague video applicants. This article has discussed these issues and gives appropriate recommendations.
1 Conference TV video coding standard review
Conference TV recognized image coding standard protocol is H series, namely ITU-T H.261, H.263 and H.264 protocols. Prior to 1995, the meeting TV system images used the H261 encoding protocol. In 1995, ITU-T developed H.263 standard for low bit rate video applications, when H263 was recognized as a pixel-based adoption of the first generation encoding technology mixed coding scheme to achieve the best results. In the next few years, ITU-T has been perfected for it to increase coding efficiency and enhance coding function. H.263 + was released in 1998; H.263 ++ was released in 2000. Although the H263 encoding technology is used to increase the compression ratio and image quality than the H261 encoding, the core of the H.263 source encoding algorithm is still the DPCM / DCT mixing encoding algorithm used in the H.261 standard, the principle block diagram Also similar to H.261.
In December 2001, ITU-T and ISO International Standardization of Video Coding Computers Configured Video Coding Group (JVT, Joint Video Team), responsible for developing a new video coding standard to achieve high compression of video Comparison, high image quality, good network adaptability and other goals. Then, the video coding standard set by JVT is defined by ITU-T as H.264; this standard is also defined as 14496-10 (MPEG-4 Part 10) High-grade video coding (AVC, Advanced Video Coding) standard.
H.264 relative to the previous encoding method, the encoding efficiency is improved in the image content prediction, using variable block size motion compensation, 1/4 sampling accuracy motion compensation, weighting prediction, etc., improve image quality, increase the error correction function And adaptability of various network environmental transmission. The test results show that the H.264 has a better PSNR performance than H.263 ++ in the medium and low bandwidth: H.264 PSNR is 3dB higher than H.263 ++. So currently the industry mainstream video conferencing system manufacturers have introduced H.264-based video products.
The H264 encoding technique enables moving image compression techniques to a higher stage, providing high quality image transfer on lower bandwidth is the application highlight of H.264. The promotion of H.264 is high for video terminal, gatekeeper, gateway, and MCU and other systems. At present, there are only real-strength manufacturers to provide products with all lines.
H.264 The coding standard is being identified by other industries, and H264 may become a unified standard for broadcast, communication, and storage media (CD, DVD), which is a unified standard for future broadband interaction new media.
2 H263 protocol or H264 comparison analysis
From the development of video standards, H.264 is a common standard determined by two major standards organizations, comprehensively covering video communication, broadcasting, storage and other applications, and the conference television system of H264 video standard protocol is reasonable. s Choice. Adoption of the H264 encoding protocol:
Macro block segmentation of different sizes and shapes: H.264 supports seven models. The minimum of 4 × 4 small pieces of small block mode improves performance, reducing the block effect, and improves the quality of the image.
High-precision sub-pixel motion compensation: The motion estimation of half a pixel precision is employed in H.263, and 1/4 or 1/8 pixel precision motion estimates can be employed in H.264. The residual residual is small after the exercise estimation.
Multi-frame prediction: H.264 provides an optional multi-frame predictive function, and 5 different reference frames can be selected when encoded, providing better error correction performance, which improves video image quality.
The deck filter: H.264 defines a filter that adaptively remove block effects, which can handle the horizontal and vertical block edges in the prediction loop, greatly reduces the square effect.
4 × 4 pieces of integer transformation: Since the size of the two transform blocks is reduced, the division of the moving object is more accurate, so that not only the conversion calculation is relatively small, but also is also greatly reduced in the end of the moving object edge;
Advanced quantization method: Unlike H.263, the fixed constant step is different, and the H.264 step is in progress with 12.5% composite rate, and the chromaticity coefficient is smaller. These measures increase the ability of the yield control and strengthen the trueity of colors.
The following is a comparison of various algorithms:
Various encoding algorithm efficiency comparison
It can be seen that the H.264 algorithm is the most leading in image codec, the MPEG-4 and H.263 algorithms are basically the same, and the MPEG-2 algorithm is low efficient, but when the image quality requires a certain degree, the bandwidth The requirements tend to normalize, that is, when high bandwidth (6m and above), the image quality provided by various coding algorithms is gradually converge.
At the same time, since H.264 codec is largely increased significantly (with the same bandwidth, the H.264 codec for the CPU is 6 times more of H.263 codec, and the current CPU processing power is difficult to support.
The judgment standard for the industry-recognized HD conference television system is: activity image is 704 × 576 resolution (4cif), 25 frames per second, considering the above factor, currently more general practice is low-bandwidth (2m or less) The H.264 encoding is used, and the H.263 encoding format is used under high bandwidth, such a match can achieve better image effect.
3 At present, construction conference television system recommendations
From the above analysis, it can be seen:
1) Since H.264 adopts higher and more effective compression technology, the equivalent bandwidth will carry more information to obtain a clearest image. In general, the H.264 of 2M bandwidth is equivalent to 3.5 M or even higher image quality of H.263 with a higher bandwidth.
2) H.264 is not only better compression algorithm, and many high-difficult new technologies, such as unified VLC symbol coding, high precision, multi-mode displacement estimation, based on 4 × 4-piece integer transformation, layered coding Syntax, deck effect filtering, inter prediction, entropy algorithm, 1/4 pixel interpolation, etc., make image information more full, the display accuracy is higher, whether it is a high bandwidth or low bandwidth network, is subject to the actual situation The impact is smaller. These technologies are unable to support other coding technologies.
3) In theory, the 8M bandwidth of H.264 is clearer than the 263 image of the 8M bandwidth, only 264 encoding needs.
It is a huge calculation that the current chip does not achieve this ability. Under normal circumstances, enterprise or industry users currently rent carriers 2M line networking, most video conferencing systems use 768K bandwidth, so 2M H.264 encoding has been better to meet existing needs, and not to stay in the future The extension is expanded.
It is recommended that the currently constructed conference television system must support H264 encoding. When the bandwidth is relatively low, the H264 encoding protocol is used to take advantage of the H.264 compressed efficiency, and the image quality is greatly improved relative to H.263; when the bandwidth recharge Adopt 4CIF's H263 encoding format.
Our other product:
