MPEG-4 is an MPEG standard ( ISO/IEC-14496 ), which describes inter alia methods for the video and audio data compression. Originally, the goal of MPEG -4 systems with low resources or narrow bandwidths was to support (mobile phone, video phone, ... ) at relatively low loss of quality. Since H.263, an ITU standard for video decoding and compression, the conditions just mentioned has been very well implemented, it has been integrated without major changes as part 2 to MPEG -4.

In addition to video decoding were also some audio standards, such as the already standardized in MPEG -2 Advanced Audio Coding (AAC ) and support for Digital Rights Management, which runs under the name IPMP ( Intellectual Property Management and Protection ) was added to the standard.

The official container format MPEG -4 MP4.

MPEG-4 Part

MPEG- 4 is composed of a plurality of standards, the "parts" (English " parts" ) will be called, including the following:

  • Part 1 (ISO / IEC 14496-1 ): Systems
  • Part 2 (ISO / IEC 14496-2 ): Visual. These include the well-known MPEG -4 codecs DivX and Xvid. One of the many profiles in Part 2 is the Advanced Simple Profile (ASP).
  • Part 3 (ISO / IEC 14496-3 ): Audio. Contains the descriptions of Advanced Audio Coding (including HE -AAC ) audio Lossless Coding, Scalable to Lossless, Direct Stream Transfer, CELP, TwinVQ and HVXC.
  • Part 4 (ISO / IEC 14496-4 ): Conformance
  • Part 5 (ISO / IEC 14496-5 ): Reference software
  • Part 6 (ISO / IEC 14496-6 ): Delivery Multimedia Integration Framework ( DMIF ).
  • Part 7 (ISO / IEC 14496-7 ): Optimized Reference Software
  • Part 8 (ISO / IEC 14496-8 ): Carriage on IP networks
  • Part 9 (ISO / IEC 14496-9 ): Reference Hardware
  • Part 10 (ISO / IEC 14496-10 ): Advanced Video Coding (AVC ). A codec for video signals which is technically identical to the H.264 standard of the ITU-T.
  • Part 11 (ISO / IEC 14496-11 ): Scene description ( BIFS ) and Application engine (MPEG -J)
  • Part 12 (ISO / IEC 14496-12 ): ISO Base Media File Format. Contains the description of the QuickTime container format.
  • Part 13 (ISO / IEC 14496-13 ): Intellectual Property Management and Protection (IPMP ) extensions
  • Part 14 (ISO / IEC 14496-14 ): MPEG -4 File Format ( MP4). Container format for MPEG-4 content based on Part 12
  • Part 15 (ISO / IEC 14496-15 ): AVC File Format. Part 10 container format for videos based on Part 12
  • Part 17 (ISO / IEC 14496-17 ): Timed Text subtitle format
  • Part 18 (ISO / IEC 14496-18 ): Font Compression and Streaming ( for OpenType fonts)
  • Part 19 (ISO / IEC 14496-19 ): Synthesized texture stream
  • Part 22 (ISO / IEC 14496-22 ): Open Font Format Specification ( OFFS ) based on OpenType
  • Part 23 (ISO / IEC 14496-23 ): Symbolic Music Representation ( SMR)
  • Part 24 (ISO / IEC 14496-24 ): Audio and systems interaction
  • Part 25 (ISO / IEC 14496-25 ): 3D Graphics Compression Model (not yet finished)

(12/ 2008).

Profiles are defined within the "parts", and therefore the implementation of a part is usually not part of the entire implementation.

Video decoding

For the Simple Profile (SP ) of the decoder part for global motion compensation can be simply omitted.

Variable length decoding

The variable-length decoding ( VLC) is based on the method of Huffman redundancy reduction. The codewords are using tables, so-called "look -up tables", decoded.

Inverse sampling

From the variable length decoding to obtain a one-dimensional array QFS [n ], there can be n values ​​from 0 to 63 to accept. This output is converted at this point in a two-dimensional array with the identifier of the block diagram above PQF [ v] [ u]. Both [v ] and [ u] ranges from 0 to 7 in the following figure, the three defined patterns are shown for sampling.

Inverse DC and AC prediction

This adaptive selection of the DC and AC Prädiktionsrichtungen based on a comparison of horizontal and vertical DC gradients around the block to be decoded around.

Inverse quantization

The quantization process itself is reversible and thus a redundancy -reducing method. However, in the reconstruction of the DCT coefficients are a limited number of discrete values. Thus, the inverse quantization is lossy. The two advantages of the quantization are:

Inverse Discrete Cosine Transform

Inverse discrete cosine transform ( IDCT) is a process for the irrelevance.

Known codec implementations ( MPEG-4 codecs)


The transport is called "Multimedia Integration Framework Delivery" in MPEG-4. The function is to hide the real source of multimedia data streams, so that the player is independent of the transport network to the terminal (possibly PC). Possible scenarios are standardized:

  • Locally ( for example, on the local hard drive )
  • Remote interactive ( client - server application, such as Video On Demand)
  • Broadcast ( unidirectional reception of data streams ).

For the local case, the file format other than "ISO Base Media File Format " in MPEG- 4 Part 12 is standardized. The company Apple introduced in 1998 the first source code. On this the MP4 file format was designed and developed the standards in part 14.

Reference software

  • Implementation Model 1 (IM1 )
  • IM1 - 2D