2.4 Advanced Video Compression Techniques
Section 2.3 discussed some of the basic video coding techniques that are common to most of the available video coding standards. This section examines some more advanced video coding techniques, which provide improved compression efficiency, additional functionality, and robustness to communication channel errors. Particular attention is paid to the H.264 video coding standard [8, 9], which is one of the most recently standardized codecs. Subsequent codecs, such as scalable H.264 and Multi-view Video Coding (MVC) , use the H.264 codec as a starting point. Note that scalability is discussed in Chapter 3.
2.4.1 Frame Types
Most modern video coding standards are able to code at least three different frame types:
- I frames (intra frames): these do not include any motion-compensated prediction from other frames. They are therefore coded completely independently of other frames. As they do not remove temporal redundancy they are usually much larger in size than other frame types. However, they are required to allow random access functionality, to prevent drift between the encoder and decoder picture buffers, and to limit the propagation of errors caused by packet loss (see Section 2.4.8).
- P frames (inter frames): these include motion-compensated prediction, and therefore remove much of the temporal redundancy in the video signal. As shown in Figure 2.9, P frames generally use a motion-compensated version of the previous frame to predict the ...