Key Technical overview and coding for MEPG-4 standards

MPEG-4 coding standards are currently the latest international coding standards specification? This article gives a brief overview of its main content? On this basis, the special audio object encoding and video object encoding are introduced.

1 multimedia transmission integration framework

The multimedia transmission integration framework (DMIF) mainly solves the Operation Problems of multimedia applications in interactive networks, broadcast environments, and disks. It integrates bit information by transmitting multiple channels, create a handshake and transmission between the client and the server. Different from the past, because the MPEG-4 code stream contains many AV objects, these AV objects generally have their own buffers, not just video buffers and audio buffers.

2 syntax description

The MPEG-4 defines a syntax description language to describe the bit stream representation and scenario description information of the AV object. This syntax description language is an extension of C ++. It is not only easy to express its AV object features, but also easy to implement in software simulation and model verification. Compared with MPEG-4, MPEG-1 and MPEG-2 use a kind of C language description, MPEG-4 Description Language reflects the object-oriented technology to describe the object.

3. Audio object Encoding

The compression encoding of video and audio is still the core of MPEG-4. However, unlike the previous MPEG-1 and MPEG-2, The MPEG-4 not only supports natural sound (such as voice and music), but also supports the synthesis of sound Based on the Description Language, supporting audio object features. That is to say, in a single scenario, people and background music may be independently encoded audio objects.

3.1 natural sound Encoding

The MPEG-4 research compares the existing Audio Encoding algorithms, supports 2 ~ 64 K natural sound encoding. For example, the sampling frequency of 8 kHz is 2 ~ 4 kbit/s Speech Encoding and 8 or 16 KHz sampling frequency 4 ~ 16 kbit/s audio encoding, generally using parameter encoding; 6 ~ 24 kbit/s Speech Encoding, generally using code Excitation Linear Prediction (CELP) encoding technology; 16 kbit/s and above bit rate encoding, can use time-frequency (T/F) transform encoding technology. These technologies actually draw on existing audio coding standards, such as g.723, g.728 and MPEG-1 and MPEG-2. Figure 1 is mpge4's scalable natural Audio Encoder, which includes three encoding technologies.

500) This. style. width = 500 ;" Onclick= Javascript: window. Open (this. SRC); src = "http://info.broadcast.hc360.com/images/tech/04122001.jpg" onLoad = "javascript: If (this. width> 500) This. style. width = 500;">

3.2 synthesize sound

In the synthesis of sound encoding, MPEG-4 introduced two very attractive coding technology: Text to Speech Encoding and music-driven synthesis coding technology. This paves the way for games that interact with voice at low bit rates on the network. In fact, the synthesis sound encoding technology is a parameter Encoding Based on the knowledge base. It is particularly worth mentioning that the music score-driven synthesis technology of the MPEG-4, in which the decoder is driven by a special synthesis language, the structured audio orchestra language (saol. The "Orchestra" is composed of different "instrument. When a decoder does not have an instrument, the MPEG-4 also allows the decoder to download the instrument from the encoder to the decoder to properly recover the synthetic sound. It can be seen that the MPEG-4 does not provide a set of angular Midi Music Standards in the "instrument", but provides a "Orchestra" that can be expanded at any time, so its "playing" Music is naturally more colorful.

4. Visual Object Encoding

Similarly, MPEG-4 also supports coding for natural and synthetic visual objects. The synthesized visual objects, such as 2D, 3D, and facial expression animations, are encoded separately, which can be effectively compressed and easy to operate.

Coding of natural visual objects is still the focus of MPEG-4. Compared with static images, MPEG-4 uses the zerotree wavelet algorithm to provide a high compression ratio while also providing up to 11 levels of spatial resolution and quality scalability.

For the coding of motion video objects, the MPEG-4 uses the block diagram shown in 2 to support image encoding.

500) This. style. width = 500 ;" Onclick= Javascript: window. Open (this. SRC); src = "http://info.broadcast.hc360.com/images/tech/04122002.jpg" onLoad = "javascript: If (this. width> 500) This. style. width = 500;">

It can be seen that in order to support object-based encoding, The MPEG-4 introduces the shape encoding module. In order to support efficient compression, MPEG-4 still uses the MPEG-1, transform in MPEG-2, prediction of mixed encoding framework.

For video objects of any shape in general, the code stream structure after MPEG-4 encoding is shown in figure 3.

500) This. style. width = 500 ;" Onclick= Javascript: window. Open (this. SRC); src = "http://info.broadcast.hc360.com/images/tech/04122003.jpg" onLoad = "javascript: If (this. width> 500) This. style. width = 500;">

For real-time Very Low Bit Rate applications, such as videophone, MPEG-4 video encoding using Very Low Bit Rate Video (vlbv) Core encoding, similar to the ITU H.263 directly on the rectangular video encoding, instead of using the shape encoding module. The encoded code stream structure is shown in Figure 4.

500) This. style. width = 500 ;" Onclick= Javascript: window. Open (this. SRC); src = "http://info.broadcast.hc360.com/images/tech/04122004.jpg" onLoad = "javascript: If (this. width> 500) This. style. width = 500;">

It can be seen that the MPEG-4 adopts forward compatible H.263, but also provides some high-level features, such as content-based encoding. See Figure 5.

500) This. style. width = 500 ;" Onclick= Javascript: window. Open (this. SRC); src = "http://info.broadcast.hc360.com/images/tech/04122005.jpg" onLoad = "javascript: If (this. width> 500) This. style. width = 500;">

The MPEG-4 supports the robustness under the error channel transmission, and provides the better synchronization and the error recovery mechanism.

5. Scenario Description

The scenario description describes how to organize and synchronize a single AV object under the coordinates of a specific AV scenario. There are also issues such as intellectual property protection for av objects and AV scenarios.

6 MPEG-4 Outlook

The Application of MPEG-4 will be extensive and far-reaching. This new standard will be applicable to at least the following scenarios:

A) real-time multimedia monitoring;

B) Mobile Multimedia Communication at an extremely low bit rate;

C) multimedia systems based on content storage and retrieval;

D) video streams and video games on the Internet/Intranet;

E) virtual conference based on facial expression simulation;

F) interactive multimedia applications on DVDs;

G) Application of Visual cooperation lab scenarios based on computer networks;

H) studio and TV program production.