| 题名 | CS-ACELP语音压缩方法的DSP实现及其应用 |
| 作者 | 汪滨 |
| 学位类别 | 博士 |
| 答辩日期 | 2004 |
| 授予单位 | 中国科学院研究生院、中国科学院声学所 |
| 授予地点 | 中国科学院研究生院、中国科学院声学所 |
| 关键词 | 语音编码 线性预测 系统实现 |
| 中文摘要 | 语音压缩编码的发展一直是在用尽可能低的数码率获得尽可能好的合成语音质量的矛盾中发展。降低数码率使得同样的信道容量能传输更多路的语音信号，或者是说，存储数字语音信号只需要较少的容量。随着数码率的降低，相应的算法延迟时间和计算复杂度也要增加。低速率语音编码技术具有重大实用价值，因此吸引着众多研究机构与公司进行研究。G．729语音编码器是国际电信联盟于1996年获准通过的采用“共扼结构代数码激励线性预测技术”的skbit／s码率的语音编码建议，其目标是在8 kbit／s码率下合成音质不低于32 kbit／s ADPCM的水平，广泛应用于多媒体通信、蜂窝移动通信、IP语音通信中。本文在介绍一些语音编码基础知识上重点研究了G.729建议的基本原理，详细分析了定点仿真软件的结构，并给出主要程序模块的流程图，最后给出一个语音实时记录系统的实现方案并在TMS320VC54O2定点DSP芯片上实时实现了该编码器过程。主要内容包括以下几个方面：1．介绍G.729用到的线性预测编码（LPC）、合成分析法（A-B-S）、线谱对分析等一些语音编码基础知识。2．研究G.729建议的算法原理。3．基于C语言编制仿真程序实现G.729语音压缩算法。4．提出一个利用G.729压缩算法实现车载语音实时记录的方案。5．介绍基于TMS32OVC54O2定点DSP芯片的语音实时记录系统的实现”方案，并给出相应的软、硬件设计方案。6．编制软件在TMS32OVC54O2 DSP上实现了G.729的语音压缩算法。7．对系统性能进行实际测试，分析实验数据及给出测试结果。 |
| 英文摘要 | The development of speech compression coding always tries to achieve the best quality of reconstrcted speech with bit rate as possible as low. Reducing bit rate makes it able to transmit more paths of speech signal at the same signal channel capacity. In other words, the less capacity will be needed for storing digital speech signal. With the decrease of bit rate, the delay of algorithm and complexity of computation relatively increase. As the technology of speech coding at low bit rates is very useful, many academic institutions and companies are preoccupied with it to have a research. The G.729 speech codec (encoder and decoder) is an algorithm recommendation for the coding of speech signals at 8 kbit/s using " Conjugate-Structure Algebraic-Code-Excited Linear-Prediction" (CS-ACELP), which was approved by ITU (International Telecommunication Union) in 1996.The aim of G.729 speech coding at 8 kbit/s is at achieving (even exceeding) the quality level of reconstructed speech using ADPCM at 32 kbit/s bit rate. ITU G.729 has been widely used in the field of multimedia communications, digital cellular communications system, IP speech communications. In this paper, the ultimatum of the recommendation is mainly studied after introducing some elementary knowledge of speech coding. The structure of ANCI C code simulating the CS-ACELP coder in 16 bit fixed-point is analyzed, and the flow chart of main software module is given. Finally the paper provides the realization scheme of a real-time voice recoding system and studies the real-time realization process of the codec based on a single chip of fixed-point TMS320VC5402 DSP. The main aspects dealt with in the paper are structured as follows: 1. Introduction of some elementary knowledge of speech coding used in G.729, including Linear-Prediction-Coding (LPC), Analysis-By-Synthesis (A-B-S), Line-Spectral-Pair and so on. 2. Studying the algorithm principle of Recommendation G.729. 3. Based on ANCI C language, programming to implement G.729 speech compression algorithm. 4.Bringing forward a scheme of implementing real-time recoding of In-car audio using Recommendation G.729 compression algorithm, 5.Introduction of a realization scheme of speech real-time recoding system, based on a fixed-point TMS320VC5402 DSP chip.Giving the corresponding implementation scheme of hardware and software. 6. Based on TMS320VC5402 DSP chip, programming to implement the speech compression algorithm of Recommendation G.729. 7. Testing the performance of the whole system, analyzing experimental data and giving the testing result. |
| 语种 | 中文 |
| 公开日期 | 2011-05-07 |
| 页码 | 73 |
| 内容类型 | 学位论文 |
| 源URL | [http://159.226.59.140/handle/311008/866]  |
| 专题 | 声学研究所_声学所博硕士学位论文_1981-2009博硕士学位论文 |
| 推荐引用方式 GB/T 7714 | 汪滨. CS-ACELP语音压缩方法的DSP实现及其应用[D]. 中国科学院研究生院、中国科学院声学所. 中国科学院研究生院、中国科学院声学所. 2004. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论