MPEG-4 - image coding standard

MPEG-4 was introduced in 1998. It is a comprehensive audio and video coding standard package including MPEG-1 and MPEG-2 coding techniques.

The standard includes new functions, e.g. VRML support for 3D rendering (3D graphics standard, interactive vector graphics, website dedicated), object oriented files including audio, video and VRML objects, DRM support (cryptography based protection system for copyright protection).

The components implemented in MPEG-4 depend on the programmer decision, and the application capabilities are determined by profiles and levels specified in ISO/IEC standards.

MPEG-4 includes general data streams which are the result of coding algorithms. The streams are sent or stored individually and can be displayed on the receiving side as an original multimedia presentation. The relation between the audio-visual elements is shown at two main levels. The first level – BIFS (Binary Format Scenes), defines time and spatial scene arrangement and allows to interact with the objects in the scene. The second level – ODs (Object Descriptors), defines the relationship between the main streams in relation to ODs and includes additional information, e.g. URL address required to access the main stream.

MPEG-4 standard in ‘Visual’ profiles allows hybrid coding of natural video sequences with virtual images created using a special computer software. The solution allows to place a person in a virtual studio (virtual reality).

MPEG-4 supports data stream from 5 kbit/s to over 1 Gbit/s in a progressive and interlaced format. It supports resolutions from sub-QCIF (88x72 px) to UHD 4k (4096×2304 px).

The compression algorithms are very efficient in compact coding of textures in high quality with a very high compression ratio. Content-based image coding allows to individually decode the selected objects in a video sequence.

Content modification is possible with the following functions: distortion, text, texture and video conversion. Coding scalability allows to vary the complexity of generated bit streams for textures or video sequences.

MPEG-4 flexibly adapts to the errors with image and video compression algorithms in error-prone environments and special tools for low quality video sequences (less than 64 kb/s), suitable for use at low bandwidths.

MPEG-4 includes an extensive set of tools for audio-visual object coding. The sets are divided into groups for different applications. The subgroups are referred to as the profiles and include different tool and coding function limits. Each profile also includes levels. A commonly used profile and level combination is 'Main Profile @ Main Level'.

New versions are always backward compatible and introduce new profiles without changing the existing profiles to maintain compatibility with previous versions.