Implementation And Optimization Of ILBC Speech Enhancement Algorithm On DSP

In modern communication systems, speech is the most important and fundamental communication way. The development of VoIP technology provides efficient means for transmission of voice service via different kinds of data network. By use of VoIP, digitalized speech signal is encapsulated in IP packet, and then the packet is transmitted over the IP network, and it is commonly transmitted in terms of bit stream by compressing. Due to various factors, such as cost, efficiency, physical channel capacity and storage space, people hope to reduce the bit-rate required for the transmission of speech signal with good speech quality. The bit-rate reduction processing is known as speech coding. Speech coding at low bit rates has been applied to VoIP. In recent years, people have paid more attention to developing high quality speech coding at low rate. So, it is a very important research issue for achieving communication quality at low bit rates.This paper used iLBC voice codec algorithm can still provide high-quality voice services in a poor network environment. Its main advantage lies in the handling capacity of packet loss. iLBC independently deals with each voice packet. It is an ideal packet-switched network voice codec. Compared with standard low-rate codecs, iLBC uses more natural, clear elements to accurately imitate the original speech signal. It's more suitable for packet-switched network and a voice codec which can obtain high-quality.This paper has in-depth study for iLBC speech codec algorithm, particular the speech enhancement module with conditions. Using the debugging tools developed by Palmmicro company achieves the uniformization of fixed-point C code and assembly code of iLBC algorithm on the line a 30ms-frame mode, that is to eliminate bugs generated in the process of artificial convert floating-point ANSI C source code to fixed-point C code and fixed-point C code to assembly code. To improve the quality of assembly code of transplantation in the DSP chip, and to optimize speech enhancement module to improve the quality of compressed voice.In conclusion testing and results analysis on the hardware platform of fixed-point DSP chip AR1688 of ADI Company shows that after fixed-point, uniformization and optimization, iLBC algorithm in engineering practice has good results, good adaptability on various network conditions, and good call quality.Voice and integrated services based on IP is an irreversible trend of history. The intervention of new technologies, new concept and new value-added service providers such as SKYPE will bring an unprecedented new era for VoIP.