MPEG2 standard compression layering and encoding principles

MPEG2 encoder can be divided into two layers: compression layer and reuse layer. Es streams are generated from the compression layer.
ES--PES, and then reuse can be PS stream or TS stream, PS is used for storage (such as DVD) ts for transmission (such as TV station playback ).

 

 

In multimedia signal transmission, high-quality sound and images can be parsed only when audio and video signals are effectively encoded. This article introduces MPEG2 standard compression layering and encoding principles, and explains how to implement MPEG2 Audio Encoding on the chip. The MPEG2 encoder provides a standard format for bit streams encoded by MPEG2. On the other hand, it also provides a standard mode for MPEG2 decoder.

 

MPEG2 is the most popular AV compression standard and can be used to store video, audio, and digital information. The complete MPEG2 standard can meet the requirements of broadcasting applications such as STB and multimedia applications such as DVDs or D-VHS. MPEG2 is not

 

The audio part of the MPEG2 standard is generally based on the mpeg1 standard, so the two are highly compatible. This allows the current mpeg1 device to decode some of mpeg1-compatible signals in MPEG2, while MPEG2 devices can also decode mpeg1 signals to achieve forward compatibility.

 

Compression layering

 

MPEG2 and mpeg1 audio compression can be divided into three layers. The higher the number of layers, the higher the compression level, the required CPU processing capacity, and the sound quality, and the less bandwidth required for transmission. Therefore, the first layer has the lowest compression rate, the lowest CPU processing capability, and the lowest latency. Since the compression rate is the lowest and the sound quality is the worst, it requires the highest transmission bandwidth. The sound quality on the third layer is the best, and the compression ratio is up. The processing time is almost three times that of the first layer. Table 1 provides a good description of this.

 

Compression is divided into three layers, one is based on needs, and the other is based on historical reasons.

 

First, it depends largely on the application to which the compressed signal will be used. For example, when we want to reproduce high-quality voices and focus on processing capability/cost, we should select Layer 3. If you mainly consider processing capabilities and costs, you can select the first or second layer. However, if you choose not the third layer, the sound will be lacking in the current review. Therefore, the user must consider the trade-off and select the layer suitable for its special applications.

 

Secondly, the concept of the compression layer is gradually developed. The third-level specification is later than the first and second-level specifications. When the third-level specification began to become popular, the devices adopting the first and second-level specifications were very popular and consumers were very familiar with it. Therefore, consumers must be free to choose the appropriate specifications for their applications.

 

MPEG2 audio compression and encoding

 

The three-layer standard audio compression and encoding process 1 is shown. Filter banks use the Fast Fourier Transform (FFT) to convert time-domain sampling to the same number of frequency-domain sampling. The output is a series of child bands with equal bandwidth. The process of acoustic mode calculates the signal masking ratio (SMR) of each sub-band to determine the number of signals that each sub-band can be used for encoding. The output of the filter group and the SMR information of each sub-band are used to determine the quantifiable noise that each sub-band can withstand. The higher the quantization noise, the lower the number of signals allocated by this sub-band. In the in-place stream formatting module, the sub-band frequency sampling is combined with the signal bit assigned to the layer and some other information to form an audio frame, this frame includes a signal header and other information segments.

 

MPEG2 audio enhancement

 

Compared with mpeg1, MPEG2 has improved in the following aspects.

 

1. Half Sampling Rate

 

In MPEG2, only half of mpeg1's sampling rate is required to maintain excellent sound quality. This is especially useful for comments channels, multilingual channels, multimedia and other applications. These channels can be used from 20Hz to 20 kHz, but are rarely used.

 

2. Multi-channel extension

 

MPEG2 supports five audio channels to achieve a "surround" stereo effect, in order to obtain a more realistic stereo. These five channels are respectively left channel (L), right channel (R), central channel (C), left rear surround channel (LS), and right rear surround channel (RS ). In this case, three tweeting speakers are arranged in front and two speakers are arranged in the back, which is also called 3/2 stereo. 2.

 

MPEG2 Audio Encoding on the chip

 

MPEG2 audio encoding can be implemented independently or together with MPEG2 Video Encoding on hardware. In the latter case, you need to add multiplexing or multi-output selector for encoding or decoding. You can also use one decoder to complete two functions. Generally, MPEG2 audio encoding is implemented through DSP, which is cheaper and more flexible. Figure 3 shows the circuit block of the MPEG2 audio decoder chip.

 

This system is available in MP3 players produced by many electronic equipment manufacturers and carries the layer-3 decoder of MPEG2 audio. Some MP3 players can store music for about 1-2 hours, which is usually stored in memory stick or flash memory. Players with hard disks can store more information.

 

To simultaneously implement MPEG2 audio and video encoding, multiplexing of the two data streams is a key issue. Audio and Video Signal Encoding can be performed on the same chip, or another chip can be used. This process is controlled by the ITU 13818-3 standard and can be selected using Program Stream or transport stream ). In the case of multiplexing, MPEG2 audio and video signals share the bandwidth. In addition to audio and video signals, the data stream also carries information about the synthesis of the two signals. In this case, the total bit rate (also known as the system bit rate) is the sum of the bit rate of the audio and video signal and the data header and byte padding. Obviously, the video part occupies most of the bandwidth. When the system bit rate is high, the bit rate of the audio signal is lower than that of the video signal or multiplexing data stream. In this case, the bandwidth is very rich, so you can set various parameters of the audio stream to get the best sound quality. However, if the bit rate is low, the situation is not the same. If the audio part wastes some bandwidth, the bandwidth of the video part will be insufficient. The final parsed video quality will be seriously impaired. In this case, you should be careful to set MPEG2 audio parameters. If "audio PES adjustment" is disabled, the audio bit rate should be reduced to 128 kbps or even kbps. These settings are very effective for Improving the Quality of MPEG2 information stream parsing.

 

Author:

Waqar Saleem

Application Engineer

Technical solution group

Fujitsu Microelectronics American company