Coding Video

4Structures

4.1 Introduction

4.1.1 How Does a Video Codec Use Structures?

A complete video clip or sequence is processed by a video encoder to create a compressed bitstream. In order to handle the large amount of image data in a sequence of video frames, the encoder splits it up into manageable structures. In this chapter, we will look at how a video encoder does this and how it bridges the gap between a source video, which contains multiple frames each made up of thousands or millions of pixels, and encoded data units. We will also look at the processing and storage capabilities of a practical codec that handles a relatively small region or block of data at a time. Figure 4.1 illustrates how a source video sequence is split up into manageable‐sized structures, such as blocks of pixels. These structures are processed and encoded to produce the compressed bitstream.

A coded video sequence is a series of coded pictures that, when decoded, will play back as a complete video programme or clip. Coded frames or pictures may be organised into multi‐picture structures during encoding, such as Group of Pictures (s). Each picture may be coded as a single unit or in multiple sections known as slices or tiles. Each slice or tile contains one or more Basic Units, such as Macroblocks or Coding Tree Units.

The Basic Unit is a unit of data handled by the encoder and decoder. In present‐day codecs, it is a square. In the older MPEG‐2 and H.264/AVC standards, it is 16 × 16 pixels, up to 64 × 64 ...