introduction
The current era is an era of rapid expansion of information. In order to meet the increasing information needs of users, various data networks and information platforms are constantly expanding capacity. however. After all, the frequency band resources are limited. Whether it is wired communication or wireless communication, the frequency band resources are becoming more and more precious. How to use reasonable data compression coding methods to save bandwidth has become a hotspot of research and application. In this paper, through the study of the compression coding method of voice signals, a set of voice junctions is finally designed, which can be connected to audio terminals such as telephones, which is very suitable for narrowband voice communication.
1 Speech coding method
From the perspective of coding method, the voice coding method can be divided into: waveform coding, parameter coding and mixed coding. Waveform coding has the advantages of strong anti-noise performance and good voice quality. But the coding rate required in this way is higher. Generally between 16 kbit / s and 64 kbit / s. The characteristic of parameter coding is the low coding rate, which can reach 1.2 kbit / s to 2.4 kbit / s. Even lower. However, the parameter encoder also has the disadvantages of poor voice quality, low naturalness, and sensitivity to the environment. Hybrid encoder, which combines the advantages of the above two methods, and constructs speech coding from two aspects: on the one hand, it increases the naturalness of the speech and improves the quality of the speech; on the other hand, it achieves a lower digital rate index compared to waveform encoding. current. Hybrid encoders are gaining greater attention. This encoder not only has the characteristics of a vocoder (using a voice generation model to extract voice parameters), but also has the characteristics of waveform coding (optimizing the excitation signal to match the input voice waveform), while also using perceptual weighting The criterion of minimum mean square error makes the encoder a closed-loop optimized system, which can obtain higher voice quality at a lower bit rate.
In this paper, the AMBE algorithm based on multi-band excitation in hybrid coding is selected. The AMBE algorithm has the following advantages: AMBE is a low bit rate, high quality improved voice compression algorithm. This technology can provide excellent performance in low bit compression systems. Speech quality, but the requirements for instruction execution speed and memory capacity are greatly reduced, it introduces new algorithms such as speech analysis and synthesis and vector quantization. It is also extremely robust in terms of background noise and channel errors. 3. 6 kb—psAMBE vocoder. The performance is equivalent to that of the full-rate (8 kbps) VSELP. Under the premise of the same voice quality, the bandwidth occupied by the AMBE coding is smaller, which saves frequency resources.
2 The composition of the voice junction system
The voice junction system mainly realizes the analog-to-digital conversion of audio data, compression and decompression, codec functions, etc., as shown in Figure 1: When the analog voice signal is converted into a digital signal after A / D conversion, the information After being compressed by the voice coding unit, it is sent to the control unit, and the information processed by the control unit can be sent out through the communication interface. When the receiver receives the information, it analyzes the data and sends it to the voice decoding unit for decompression. Then the decompressed digital information undergoes D / A conversion to generate an analog voice signal and send it to the voice terminal.
3 Voice codec chip selection and interface circuit design
In the system design, the A / D and D / A chips select the voice band CodecCSPl027, which is suitable for the development of cellular phone and modem applications launched by LHeent. CSPl027 integrates 16-bit ∑A / D, D / A on the chip. Connect with the subsequent system through 16-bit serial I / 0 port. CSPl027 has a certain software programming control function, according to different application requirements. Through software programming to control the gain attenuation, sampling rate and interface mode of the audio interface of the system. This design improves the overall integration of the system, enhances its reliability, and satisfies the system's design requirements for voice channels.
The voice compression uses the domestic low bit rate vocoder AMBE-1000, which is a high-performance voice codec based on the multi-band excitation concept AMBE technology successfully applied, the coding quality is significantly better than CELP, RELP, VSELP, MELP, ECELP, MP-MLQ, LPC-lO and other coding schemes have high synthetic speech quality and strong ability to resist background noise and bit errors. The encoding rate is adjustable from 2.4 kbPs to 9.6 kbPs, and the FEC rate can be set online from 50 bps to 7.2 kbps. The main processor of the module can determine the bit error rate according to the prior formula The strategy adjusts the voice coding rate and FEC rate to maintain the best voice quality as possible. When the channel quality is very good, the system can appropriately reduce the voice coding rate and FEC data rate to transmit more voice signals and increase the channel multiplexing efficiency; when the channel quality is poor, the system encodes the voice coding rate and FEC The data rate can be increased appropriately. To ensure a certain transmission quality. This system can achieve related rate combinations.
Siemens Resin Cast (Dry) Distribution Transformer
Switchgear Circuit Breaker,Sudhir Switchgear Junction Box,Generator Control Switchgear,Middle Voltage Switchgear
Shandong Shunkai electrical equipment co., LTD. , https://www.chinasdsk.com