面罩语音质量评价算法适用性研究

发布时间：2018-07-03 21:09

本文选题：面罩语音 + 维纳滤波　；参考：《计算机工程与应用》2017年19期

【摘要】：针对语音编码的音质评价算法性能已十分明确,但对于面罩语音不一定适用。讨论了语音质量评价算法对空气语音与面罩语音在不同噪声环境下的适用性。采用主观意见得分和三种客观评价测度对多种信噪比的带噪语音和增强语音进行评价,包括分段信噪比、改进的巴克谱失真(MBSD)和语音感知质量评价(PESQ),根据与主观评价的一致性判断客观评价方法的适用性。增强算法采用维纳滤波法和对数谱最小均方误差法(LSA-MMSE),噪声采用粉红噪声、海浪噪声。仿真结果表明,语音质量评价算法的适用性与语音类型、信噪比、背景噪声、增强算法种类有关。粉红噪声环境下,PESQ不适合评价经维纳滤波增强的空气语音;MBSD算法只适用于评价经LSA-MMSE增强的面罩语音。海浪噪声环境下,PESQ适用于评价面罩语音,MBSD不适合评价面罩语音。
[Abstract]:The performance of speech quality evaluation algorithm for speech coding is very clear, but it is not suitable for mask speech. The applicability of speech quality evaluation algorithm to air speech and mask speech in different noise environments is discussed. Subjective opinion scores and three objective evaluation measures are used to evaluate noisy speech and enhanced speech, including segmented signal-to-noise ratio (SNR). The improved Barker Spectral distortion (MBSD) and speech perception quality Assessment (PESQ) are used to judge the applicability of the objective evaluation method according to the consistency with the subjective evaluation. The enhancement algorithm uses Wiener filter and LSA-MMSE (logarithmic spectrum least mean square error) method, and the noise is pink noise and wave noise. Simulation results show that the applicability of speech quality evaluation algorithm is related to speech type, SNR, background noise and enhancement algorithm. PESQ is not suitable for evaluating air speech enhanced by Wiener filter and MBSD algorithm is only suitable for evaluating mask speech enhanced by LSA-MMSE. PESQ is suitable for the evaluation of mask speech and MBSD is not suitable for evaluation of mask speech.
【作者单位】：河北工业大学电子信息工程学院;天津商业大学信息工程学院;
【基金】：天津市自然科学基金重点项目(No.14JCZDJC32600)
【分类号】：TN912.3

【相似文献】