H264 in the I / B / P frames
- I frame is a key frame, it uses intraframe compression techniques
- Before and after the B-frame is a reference frame, which belongs to the inter-frame compression techniques, prior to compression B frames, frames before and after the non-compressed video will reference, both before and after the recording is not stored frames "residuals", in order to get better compression ratio
- Backward reference frame is a P frame, P frame also belongs to interframe compression techniques for B frames, the P frame compression rate lower than the frame b
In order to reduce latency live, real-time interaction when to use less B frames need to cache more data at the time of compression and decoding of B frames
GOF
- A is a set of frames GOF, group of frames may comprise I frames and P-frames and B
SPS and PPS (receive before a group of frames, must be received)
- SPS: sequence Parameter Set, a sequence parameter set, storing the number of frames, number of reference frames, the decoded image size, frame identification field coding mode selection
- PPS: picture Paramter Set, level image parameters, entropy coding mode select identification stored, the number of slice groups, and the initial quantization parameter identification deblocking filter coefficient adjusting
Video Huaping / Caton reason
- If a P frame GOP packet loss will be adhered to the image decoding side error
- To avoid the problems of the video, it is generally found that if the P frame or I frame is lost, which is not displayed GOP all frames until the next refresh the image after the I frame to
Video What video codec
- X264 / X265 (higher compression ratio, direct seeding system is not recommended)
- openH264: SVC support, performance may be lower than an X264 data is hierarchically distributed
- VP8/VP9