一种新的语音信号波形编码的改进算法研究
文献类型:学位论文
作者 | 宋岩涛 |
学位类别 | 博士 |
答辩日期 | 1998 |
授予单位 | 中国科学院声学所 |
授予地点 | 中国科学院声学所 |
关键词 | 语音信号时域半波波形 矢量量化(VQ) 语音压缩编码 |
中文摘要 | 本文主要阐述了一种语音信号波形压缩编码算法。该算法根据语音信号本身的特点,在对语音信号进行编码时,将语音信号分成三个部分:无音段,清音段,浊音段,分别对三部分进行编码。这三个语音信号部分都有其本身特点,我们根据三部分的不同特点,分配不同的码率。对无音段,以一个连续无音段作为一个整体进行编码,分配给极少的码字来表示它;对于清音段,以固定帧长进行量化压缩;对于浊音段,对语音信号时域上的半波进行矢量量化,以达到压缩编码的目的。本文介绍了算法的具体实施,码本的生成,算法性能的研究,以及一些背景知识的介绍。同时也着重介绍了我对这个算法的所采用的改进之处。在本文末尾还介绍了算法中存在的问题及可能的改进方案。本文所介绍的语音压缩编码算法具有运算量小,压缩比高,码率低,恢复后的语音质量较好等特点。经过我们测试和计算,在11kHz采样频率,16bit量化的情况下,浊音部分,清音部分,清浊联合部分的平均码率分别为4bit/sample,0.591 bit/sample,1.18 bit/sample,平均压缩比分别为4,27.1,13.6倍;无音段的压缩比与无音段的长度成正比,它的码率和有音部分的码率相比可以忽略不计。从中我们可以看出这个算法具有大压缩比和低码率,在不远的将来可以被广泛应用于语音传输和存储等领域。 |
英文摘要 | In this paper we mainly discuss a new kind of Speech Signal Compression Coding Algorithm based on Waveform. According to the features of Speech Signal, before we encode the Speech Signal, we segment Speech Signal into three parts: Silence segment, Unvoiced sound segment, Voiced sound segment. According to the different features of the segments, we allocate different bit rate to each type of segment. We encode a continuous silence segment as an encoded data, allocate a little bit rate to the silence segment. We take the Unvoiced sound segment as a series of frames that have fixed length, and encode it. For the Voiced sound segment, we take half-waveform of Speech Signal on time domain as the encoded unit. This paper introduces procedures of the algorithm, how to generate the codebook, the research about the algorithm performance, and some basic knowledge about Speech Signal Processing. At the meantime, we emphasize the works I did to this algorithm. At the end of this paper, we discuss the problems in this algorithm and how to resolve the problems. The features of this Speech Signal Compression Coding Algorithm are: low computing time, high compression ratio, low bit rate etc. According to tests or calculation, the mean bit rate of Voiced Sound Segment, Unvoiced sound segment, the Voice and Unvoiced segment are about: 4 bit/sample, 0.591 bit/sample, 1.18bit/sample; the mean compression ratio are about: 4, 27,1, 13.6; the bit rate of Silence segment is very little, can be ignored. From the above, we can know that the algorithm have high compression ratio and low bit rate, so it can be used diffusely in Speech Signal Communication and restoring in near future.(图版 24个, 表格 2个, 参考文献 25个) |
语种 | 中文 |
公开日期 | 2011-05-07 |
页码 | 44 |
源URL | [http://159.226.59.140/handle/311008/1386] ![]() |
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 | 宋岩涛. 一种新的语音信号波形编码的改进算法研究[D]. 中国科学院声学所. 中国科学院声学所. 1998. |
入库方式: OAI收割
来源:声学研究所
浏览0
下载0
收藏0
其他版本
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。