Font Size: a A A

The Research And DSP Realization On G.729 Voice Compression Coding Algorithm

Posted on:2015-11-16Degree:MasterType:Thesis
Country:ChinaCandidate:J WangFull Text:PDF
GTID:2308330473451751Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
G.729 speech coding standard is proposed by the International Telecommunication Union. It is widely used in voice communication, because of a lower coding rate, low time delay and a higher quality of the synthesized speech.At present, the G.729 speech coding algorithm has been very mature, but it still has some problems: First of all, G.729 speech coding algorithm complexity is high, for speech communications, the key factor is real-time, and the existing standard algorithm is difficult to be used directly. Secondly, although the G.729 standard has been proposed for many years, the integration of the G.729 codec is relatively little. In this respect, according to G.729 standard and the handheld information platform which is researched by the group independent to research and implement in this thesis.Firstly, the thesis researches G.729 speech coding algorithm. It analyses and simulates the G.729 standard algorithm to get the codec complexity of the algorithm in order to provide the basis of real-time encoding and decoding.Secondly, the thesis designs DSP software architecture of the audio codec algorithm. It design audio capture and playback module, voice codec module, interrupt service routine module and optimize the program from compiler level and code level, which is based on DSP kernel in DM3730 dual-core chip and audio chip to ensure that G.729 codec algorithm can process speech signals real-time.Thirdly, the thesis verifies the performance of G.729 codec algorithm. It tests audio capture and playback module and audio codec module which are through DSP online debugging tools base on the handheld information platform. We can get the average segment SNR of the synthetic speech quality is 10 dB, coding a frame speech signal requires 5.4ms and decoding a frame speech signal requires 1.7ms, which is meets real-time requirements and achieves the expected goals.In this thesis, it offers a solution for designing and implementing of speech signal capture and codec based on DSP and it has practical value.
Keywords/Search Tags:G.729, Speech coding, CODEC, DSP
PDF Full Text Request
Related items