History and characteristics of the MPEG standards. Part 3
3) The MPEG-4 standard is a special article. MPEG-4 is not just an algorithm for compressing, storing and transmitting video or audio information. MPEG-4 is a new way of presenting information, it is an object-oriented representation of multimedia data. The standard operates with objects, organizes hierarchies, classes, etc. from them, he builds scenes and controls their transfer.
The objects can be ordinary audio or video streams, as well as synthesized audio and graphics data (voice, text, effects, sounds …). These scenes are described in a special language. We will not dwell on this standard in detail; this is a topic for a separate extensive discussion. It can only be said that as a means of audio compression in MPEG-4, a set of various audio coding standards is used: the MPEG-2 AAC algorithm, the TwinVQ algorithm, as well as HVXC (Excitation Coding) voice coding algorithms. harmonic vector) – for 2-4 Kbps bit rates and CELP (Code Excited Linear Predictive) – for 4-24 Kbps bit rates. In addition, MPEG-4 has many scalability mechanisms.
4) The MPEG-7 standard, the development of which has not yet been completed, is fundamentally different from all other MPEG standards. The standard is not being developed to establish a framework for transferring data or writing and describing data of any particular kind. The standard is intended to be descriptive, intended to regulate the characteristics of any type of data, even analog. The use of MPEG-7 is intended to be closely related to MPEG-4. MPEG-7 is scheduled for release in 2001.
For the convenience of handling compressed streams, all MPEG algorithms are designed in such a way that they allow decompression (retrieval) and playback of a stream simultaneously with its reception (download) – stream decompression “on the fly” (stream playback) . This opportunity is widely used on the Internet, where the speed of information transfer is limited, and with the use of these algorithms, it is possible to process the information at the moment it is received without waiting for the end of the transfer.
What are CBR and VBR?
As you know, the result of encoding a signal using an algorithm such as MPEG-1 Layer III (MP3) (or some other algorithms) is a bit stream with a frame (block) structure. This is due to the fact that the source stream is not encoded in its entirety, but in parts. That is, in fact, the original stream is divided into blocks of a certain fixed length, then each block (frame) is encoded individually, and the result (encoded information block) is sent to the resulting stream (either a file or a stream of data).
CBR (constant bit rate) is a method of encoding the original audio stream, in which all its blocks (frames) are encoded with the same parameters (with the same bit rate). In other words, the bitrate over the entire length (all frames) of the resulting stream is constant.
VBR (Variable Bit Rate) is a method of encoding the original audio stream, in which each separate block (frame) is encoded with its own bit rate. The choice of the optimal bit rate to encode a given frame is made by the encoder itself by analyzing the “signal complexity” in each individual frame.
