摘要
为有效解决现有语音信号压缩算法基本没有考虑人耳听觉特性,所划分出的子带与人耳听觉特性相去甚远,语音质量不高的问题,提出了一种基于不完全小波包分解的语音数据压缩算法。该算法充分考虑语音信号的特点以及人耳听觉特性,利用小波包合理分割子带,在每个子带内进行编码,并采用优化目标函数,作为选择最优小波基的评价函数,使划分出的子带更符合人耳听觉特性。实例仿真计算表明,该方法能取得较高的压缩率,压缩后恢复的语音信号具有良好的清晰度和自然度。
One kind of speech signal data compression algorithm based on non-complete wavelet packet resolution is proposed; to effectively solve the problems that the present speech signal compression algorithms did not considered basically the characteristic of human hearing, and the divided sub-band deviated away from human hearing very much, and the speech quality is not high. This algorithm considered fully the speech signal characteristic as well as the characteristic of human hearing, and divided reasouably the sub-band with wavelet packet and coded in each sub-band. The optimum goal function as the optimum wavelet base appraisal function so as to the result is more close to human hearing. The example simulated results indicated that this method had high data compression ratio and the reconstructed speech signals had good quality.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第10期2471-2474,共4页
Computer Engineering and Design
基金
上海市教委科技发展基金项目(050Z02)
校科技基金项目(KJ2009-12)
关键词
数据压缩
小波包变换
语音信号
目标函数
压缩率
data compression
wavelet packet transform
speech signal
goal function
compression ratio
