Key words:
Code, digital TV
A typical digital television system is a source of digital television program (mainly composed of data, audio, and other data). Get compressed encoded video, audio code stream, and subsequent source encoding, which requires support for data and control data. Channel coding implementation error error, error correction, to improve the anti-interference capability of digital television transmission signals, so that the channel transmission characteristics are adapted, and carrier modulation is performed to implement spectrum movement. At present, there are two major organizations that develop video coding standards, which are ITU-T and ISO / IEC. The standards set by ITU-T include H. 261, H. 263, h. 264, mainly used in real-time video communications, such as conference TV. The MPEG series standard is made by ISO / IEC, primarily applied to video storage (VCD, DVD), streaming, radio television, Internet, or wireless online streaming. H. 264 has a strong anti-chance characteristic, adapted to a high packet loss rate, interrupted video transmission in a serious wireless channel.
1 Code technology analysis
1.1 intra predictive coding
The intra encoding is used to reduce the space redundancy of the image. In order to improve H. The efficiency of the 264 intra coding is sufficiently utilized in a given frame, and the spatial correlation of the neighboring macroblock is taken, and the adjacent macroblocks typically contain similar properties. Therefore, when encoding a given macroblock, first, according to the surrounding macroblock prediction (typically based on the macroblock in the upper left corner, because the macroblock has been encoded), then the predicted value and the actual value The value is encoded so that the code rate can be greatly reduced relative to the frame encoding directly. Figure 1.
1.2 Inter-frame prediction code
Inter-frame prediction coding utilizes time redundancy in the continuous frame to perform motion estimation and compensation. H. 264 Motion compensation supports most of the key features in previous video coding standards.
(1) macroblock segmentation of different sizes and shapes
Different sizes and shapes can be used for motion compensation for each 16 × 16 pixel macroblock.
(2) High-precision sub-pixel motion compensation
H. The 263 is used is the motion estimate of the half pixel precision, and in H. The motion valuation of 1/4 or 1/8 pixel precision can be employed in 264. In the case where the same accuracy is required, h. 264 Use 1/4 or 1/8 pixel precision of the residual to the H. 263 The residual after the estimation of the half pixel precision movement is small. This is in the same accuracy, h. 264 The code rate required in the frame encoding is smaller.
(3) Multi-frame prediction
H. 264 provides an optional multi-frame prediction feature, and 5 different reference frames can be selected when encoded, which provides better error correction performance, which can improve video image quality. This feature is primarily applied to the following occasions: cyclical motion, translation, transform the camera in two different scenes.
(4) Go to block filters
H. 264 Defines a filter that adaptively remove block effects, which can handle the horizontal and vertical block edges in the prediction loop, greatly reduce square effects.
1.3 entropy encoding
The final step of video encoding processing is entropy encoding, in H. There are three entropy encoding methods in 264. The first is an exp-golomb code that is suitable for all other factors other than the transform coefficient. The second is based on context-based adaptive beam coding (CAVLC), which is suitable for transform coefficients. The third is based on context-based adaptive binary arithmetic coding (CABAC). The CAV-LC is encoded by the CABAC based on the neighboring block of the neighboring block to achieve better coding efficiency. CABAC is higher than the CAVLC compression ratio, but it is more complicated.
(1) EXP-GOLOMB code: The simplest entropy encoding method is to all syntax elements, in addition to the quantization coefficient, use a single unlimited scalable codeword table, so you don't have to design a dedicated VLC table for each syntax element. . It is suitable for all symbols other than transform coefficients such as syntax elements. It is a variable length code (such as Table 1) with a rule structure.
(2) CAVLC encoding: a method of encoding a 4 × 4 (and 2 × 2) block transform coefficient used to zig-ZAG scanning of the difference to the difference. CAVLC utilizes several characteristics of quantified 4 × 4 blocks: 1 predicting, transformation and denial of blocks are generally sparse (containing a lot of 0). CAVLC is compactly representing a long string zip using a run. The highest non-zero coefficient after the 2zig-ZAG scan is often the sequence of + 1 / -1. 3 The number of non-zero coefficients of adjacent blocks is related. The number of coefficients is encoded with a comparison table, and the selection of the control table depends on the number of non-zero coefficients in adjacent blocks. 4 The amplitude of the non-zero coefficient is relatively high in the beginning of the rearrangement array (near DC coefficient), is relatively low in the high frequency. CAVLC uses this to appropriately select the VLC comparison table of amplitude parameters according to the most recent encoded amplitude.
(3) CABAC method: Arithmetic encoding enables the probability model of all syntax elements (transform coefficients, motion vectors) on both sides of the encoding and decoding. In order to improve the efficiency of the arithmetic coding, the basic probability model can be adapted to the statistical characteristics change with the video frame by the process of modeling. Content modeling provides the conditional probability estimation of encoding symbols, using the appropriate content model, the correlation between the symbols can be removed by selecting the corresponding probability model of the encoded symbols currently to encode symbols, and different syntax elements are usually maintained. Different models. It is understood that the basic steps of the context-based adaptive binary arithmetic coding algorithm are: 1 for each syntax element, select the probability model according to its context. 2 Based on local statistics adaptive adjustment probability statistics 3 use arithmetic coding. The block diagram is shown in Figure 2.
Context Probability Selection: Select "Context Site" based on past observations. This model is a probability model of one or several binary symbols, which can be selected from a valid model that is based on the statistics of the data symbols of the most recent priced. The context model stores each binary symbol is probability of 1 or 0. Binaryization: Convert non-binary symbols (transform coefficients or motion vectors) into binary code. ABAC only processes 1 and 0.
Arithmetic Coding Engine: A probability model selected by an arithmetic encoder encodes each bit. There are only two sub-intervals (equivalent to 0 and 1) probability update for each bit: the selected context model is updated according to the actual encoded value (if the bit value is "1", the overview of the frequency "1") .
2 H. 264 Application in video conferencing
At present, most video conferencing systems are H. 261 or H. 263 video coding standard, and H. The appearance of 264 makes h. 264 can be compared to H. 263 Reduce the rate of 50%. That is, the user can enjoy H. Even with bandwidth of 384kbit / s. High quality video services up to 768kbit / s at 263. H. 264 Not only helps save huge expenses, but also improve the efficiency of resources, while making the video conferencing services to the commercial quality have more potential customers.
IP and wireless environment H. The 264 draft contains tools for erroneous elimination, which makes it easy to compress the video in the error, packet multi-environments, such as the robustness that is transmitted in the mobile channel or IP channel.
In order to defend the transmission error, h. The time synchronization in the 264 video stream can be completed by using the intra image refresh, and the space synchronization is supported by the SLICESTRURED CODING. At the same time, in order to facilitate the error of the error, a certain focal point is also provided in video data of an image. In addition, the intra macroblock refresh and multi-reference macroblock allows the encoder to not only coding efficiency while determining the macroblock mode, but also consider the characteristics of the transport channel. In addition to the use of quantitative steps to adapt to the channel code rate, in H. In 264, the method of data segmentation is often used to deal with the change in channel code rate. Overall, the concept of data segmentation is to generate video data with different priority to support service quality QoS in the encoder. For example, the syntax-based data partitioning method is used, and each frame data is divided into several parts, which allows not discarding less important information when the buffer overflows. Simpore Data Partitioning methods can also be used to accomplish by using multiple reference frames in the P frame and B frames.
In the application of wireless communication, we can support the bit rate change of the wireless channel by changing the quantization accuracy or spatial resolution of each frame. However, in the case of multicast, it is impossible to respond to various bit rates of the encoder. Therefore, different from the fine classification of FGS (Fine Granular Scalability) used in MPEG-4 (low efficiency is relatively low), H. 264 SP frames are used to replace the hierarchical encoding.
3 Conclusion
Therefore, the video conferencing product pays the H.264 protocol, and researchers are committed to popularizing the new industry standards in H.264. We will experience H.264 video services to bring us high quality enjoyment.
Our other product:
