语音识别中带宽失配的补偿研究  

Research on Bandwidth Mismatch Compensation in Speech Recognition

作　　者：何勇军[1,2] 韩纪庆[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001 [2]哈尔滨理工大学计算机科学与技术学院,哈尔滨150080

出　　处：《计算机学报》2011年第9期1629-1637,共9页Chinese Journal of Computers

基　　金：国家"八六三"高技术研究发展计划项目基金(2006AA010103);国家"九七三"重点基础研究发展规划项目基金(2007CB311100)资助~~

摘　　要：目前的语音识别系统在训练环境与测试环境匹配的情况下具有很高的识别率,而当环境失配时,其性能将急剧下降.作者研究发现,带宽失配,即训练语料和测试语料带宽不一致,也是引起环境失配的主要原因之一.当测试语音带宽比训练语音带宽窄时,丢失的频段不可逆,且其影响在倒谱域或对数频谱域上是时变的,因而无法用目前的信道补偿方法补偿.文章在分析丢失频段对梅尔频率倒谱系数影响的基础上,提出了用频谱折叠方法对窄带测试语音进行补偿.在此基础上给出了语音带宽检测算法和带宽补偿统一框架.在AN4和TIMIT/NTIMIT数据库上的实验表明,该框架能有效增强语音识别系统在带宽失配情况下的鲁棒性.Speech recognition systems obtaining high recognition rates in clean environments perform badly in mismatch environments without compensation. Based on the research, we found that bandwidth mismatch, namely the bandwidth difference between the training and test conditions, is one of the main factors leading to environment mismatch. When the bandwidth of the test speech is narrower than that of the training speech, the distortion is non-invertible and timevarying in the logarithm spectrum and cepstrurn domains. So it could not be compensated with current channel compensation methods. After analyzing the Mel-frequency cepstrum coefficient distortion caused by the lost frequency band, we propose a compensation method based on spectral fold. Furthermore, we provide an algorithm for speech bandwidth detection and a unified compensation framework. Experiments on the AN4 and TIMIT/TIMIT databases show that the proposed framework improved the robustness of speech recognition underbandwidth mismatch conditions.

关 键 词：带宽失配 畸变补偿 梅尔倒谱 鲁棒性 语音识别 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]