MPEG-2 Video
MPEG-2 Video, standardized as ITU-T H.262 and ISO/IEC 13818-2, is a compressed video format found in older disc and broadcast media.
An older television recording or disc-derived film often carries MPEG-2 Video, standardized as ITU-T H.262 and ISO/IEC 13818-2. Its compressed track describes how to build pictures; a transport or program stream supplies the packaging around that track.
To understand the format, follow one of its coded pictures.
Sequence headers, macroblocks, and DCT corrections
Sequence information establishes the picture grid and relevant timing/format parameters. Picture information identifies the coding type and applicable reconstruction choices.
In common 8-bit 4:2:0 material, a macroblock covers a 16 × 16 brightness region. Its brightness corrections are represented in four 8 × 8 blocks, with corresponding color-difference blocks.
The encoder uses a discrete cosine transform (DCT) to represent correction patterns as coefficients. Quantization reduces their precision. The decoder reconstructs those coefficients, applies the inverse transform, and adds the correction to the picture's predictor.
The block representation explains familiar low-bitrate artifacts: lost fine patterns or disagreeing coded regions can reveal softened texture and block-shaped boundaries.
Picture dependencies and decoding order
An I-picture reconstructs its picture content without a temporal sample reference. A P-picture can predict from an available anchor and encode what changed. A B-picture can use the permitted prediction from surrounding anchors, which have to be decoded first.
For a visible sequence I0, B1, B2, P3, the decoder can reconstruct I0 and P3 before B1 and B2. It then shows them in their original viewing order.
MPEG-2's I/P anchor structure differs from modern structures that can use B-pictures as references. The GOP explanation gives the general dependency trace; the specific codec's rules determine the legal graph.
An entry at a GOP boundary also needs appropriate access handling if picture dependencies extend across it.
Interpret a broadcast picture's time
MPEG-2 can represent progressive and interlaced material. A reported 25 fps recording can mean 25 field pairs containing 50 native field times/s, rather than 25 complete progressive camera moments.
Its picture and sequence signaling guide field presentation. Film transfers can instead use repeated-field timing to place a progressive source on a television cadence. The reconstruction task is then cadence recovery, while native interlaced movement calls for deinterlacing if the output is progressive.
These differences matter when creating a new playback version: preserving a film's picture sequence and preserving a sport recording's field movement are different jobs.
Identify the surrounding names
| Name | Place in the recording |
|---|---|
| MPEG-2 Video / H.262 | Picture compression |
| MPEG Transport Stream | Packetized carriage of the video and other streams |
| MPEG Program Stream | Another systems packaging form |
| MPEG Audio | Separate soundtrack coding |
A TS file can carry video codecs other than MPEG-2, and a MPEG-2 picture track can be carried in other supported packaging.