Differences between mpeg2, mpeg4, and h264 encoding standards

1. Macro Block matching pixel precision: In MPEG2, the motion estimation accuracy is 12 pixels, which can be implemented through linear insertion (which may be simply corrected ); both H264 and MPEG4 support 14-pixel precision. 2. Number of reference frames: MPEG2. P frames of MPEG4 can only have one reference frame, and B frames can have two reference frames. H. 264 can have up to 15 reference frames, one

1. Macro Block matching pixel precision: In MPEG2, the motion estimation accuracy is 1/2 pixels, which can be implemented through linear insertion (which may be simply corrected ); both H264 and MPEG4 support 1/4 pixel precision. 2. Number of reference frames: MPEG2. P frames of MPEG4 can only have one reference frame, and B frames can have two reference frames. H. 264 can have up to 15 reference frames, one

1. Macro Block matching pixel precision: In MPEG2, the motion estimation accuracy is 1/2 pixels, which is achieved through linear interpolation (which may be simply corrected ); both H264 and MPEG4 support 1/4 Pixel Precision

2. Number of reference frames: MPEG2, MPEG4 P frames can only have one reference frame, B frame can have two reference frames; H. 264 you can have up to 15 reference frames. Generally, 3-5 frames are selected as the reference frames.

3. Filter Structure: The interpolation filters are different. In MPEG4, simple linear interpolation is used, while in H264, the 5-stroke filter is used. MPEG2 does not have the deblocking tool in the ring, and MPEG4 has the deblocking tool, but it belongs to postprocessing; H264 has the deblocking tool in the ring.

4. I-frame encoding: MPEG2 does not have INTRA prediction, but has DC parameter prediction. MPEG4 does not have INTRA prediction, supports DC prediction, and supports prediction of Two AC parameters. H264 supports INTRA prediction. There are 9 modes in 4*4.

5. Differences in macro block size matching: the minimum size of MPEG2 is 8*8, and MPEG4 is 8*8. The minimum size of H.264 can be 4*4.

6. Differences in DCT Transformation: the MPEG series are 8x8 DCT transformations, which may cause drift during Reverse conversion. The H264 is an integer 4x4 DCT transformation, in addition, the DC coefficient is transformed by DCT after HADMARD transformation, and the integer DCT transformation can prevent drift during Reverse conversion.

7. Quantified encoding: MPEG2/MPEG4 is based on huffman, And the header syntax is relatively simple, so variable length codes are rarely used. H264 provides context-based CAVLC and CABAC.

8. Encoded structure: MPEG2/MPEG4 is a hierarchical structure, while H264 provides a single-layer structure of NAL and a set of parameters.

9. Differences between quantization parameters: MPEG4 provides 31 quantization parameters (quantization mode: constant, default quantization matrix, and custom quantization matrix ), h264 provides 51 quantization parameters (quantization mode: constant, quantization matrix, whether the user-defined parameters are uncertain );