语音压缩编码技术是数字通信技术中非常重要的部分。随着通信、计算机网络等技术的飞速发展，语音压缩编码技术得到了快速发展和广泛应用。CELP编码是上世纪八十年代发展起来的一种有效的语音压缩编码方案，现已广泛地应用于集群通信、多媒体通信以及IP电话通信中。 G.729协议是CELP编码方案的一个特例，它是ITU-T推荐的语音压缩编码中最复杂的一种，它使用了当前语音压缩编码的各种先进技术，计算量较大，但是其编码合成语音质量较高，具有很高的研究价值。 本文首先对G.729协议的编解码方案及协议中应用的关键语音编码技术进行了剖析，在对G.729协议编码方案进行深入研究的基础上，对矢量量化算法进行了优化，减少了矢量码本搜索的复杂度，加快了矢量码本搜索速度；对固定码本搜索算法进行了改进，提出了一种自适应子矢量共轭结构代数码本激励算法，该算法保持了G.729共轭结构代数码本的优点，使编码传输速率从原来的8kbps降低倒7.2kbps，压缩了传输带宽，更重要的是，它的算法复杂度仅为原G.729固定码本搜索算法复杂度的1/2左右；最后，本文用MATLAB搭建了编码器仿真平台，对改进算法进行了验证，仿真结果表明，改进后的编码器在降低算法复杂度和传输速率的前提下，译码合成语音质量仍然较高，具有较好的可懂度、清晰度和自然度。

Audio compression encoding technology is a very important part in the field of digital communication. With the progress of communication and computer network, audio compression encoding technology has got a rapid development and broadly used in many places .CELP speech encoding technology was developed from the late of last century, and now has broadly used in trunking communication , multimedia communication and IP telecom. G.729 agreement is a special implementation of CELP. It is the most complicated audio compression encoding protocol that ITU-T has ever recommended. G.729 imported many prevailing encoding technology. It takes a great lot of calculation, but the quality of its reconstructed speech is outstanding which makes it really deserved to research on. In this paper, several key encoding methods of G.729 protocol were analyzed at first, on the basis of deep research of the agreement, a normal verdict algebraic was improved to optimize the vector quantization method which decreased the search complexity of the vector codebook and fastened the search velocity. A new fixed codebook exiting method—Adaptive sub-vector conjugate algebraic structure codebook exiting method was proposed which inherited the outstanding quality of the conjugate algebraic structure codebook. The rate of the improved codec is 7.2kbps ，better than the former 8kbps. And what is more important is the computation complexity of the new algebraic is only one second of the former. In order to proof the feasibility and quality of the improved algebraic, a simulation workbench was carried out with MATLAB. The simulation result indicated that the improved codec is preferable with comprehensible, clear and nature reconstructed speech on the decreased computation complexity and bit rate.