论文标题
自动估计语音中辅音的清晰度度量
Automatic Estimation of Intelligibility Measure for Consonants in Speech
论文作者
论文摘要
在本文中，我们提供了一个模型，以估算单个语音细分的可理解性的真实评估度量。我们训练了基于卷积神经网络（CNN）的回归模型，以进行停止辅音\ textipa {/p，t，k，b，b，d，d，g/}与元音\ textipa {/a/}关联的相关的回归辅音，以估计相应的信号与噪声比（SNR）的声音（CV）声音（cv）声音（cv）的声音（n Normant ears n n n n n n n n n n n n n n n n n n n n.每种声音的可理解性度量称为SNR $ _ {90} $，被定义为人类参与者能够平均至少正确识别90 \％的SNR级别，如先前对NH受试者的实验所确定的。将CNN的性能与基于自动语音识别（ASR）的基线预测进行了比较，具体来说，从SNR中减去ASR能够正确标记辅音的恒定偏移。与基线相比，我们的模型能够准确估计SNR $ _ {90} $〜清晰度度量，平均平均值（MSE）少于2 [db $^2 $]，而基线ASR定义的度量计算SNR $ _ {90} $〜，差异为5.2至26.6 [db $^2 $^2 $^2 $^2 $]，consection。
In this article, we provide a model to estimate a real-valued measure of the intelligibility of individual speech segments. We trained regression models based on Convolutional Neural Networks (CNN) for stop consonants \textipa{/p,t,k,b,d,g/} associated with vowel \textipa{/A/}, to estimate the corresponding Signal to Noise Ratio (SNR) at which the Consonant-Vowel (CV) sound becomes intelligible for Normal Hearing (NH) ears. The intelligibility measure for each sound is called SNR$_{90}$, and is defined to be the SNR level at which human participants are able to recognize the consonant at least 90\% correctly, on average, as determined in prior experiments with NH subjects. Performance of the CNN is compared to a baseline prediction based on automatic speech recognition (ASR), specifically, a constant offset subtracted from the SNR at which the ASR becomes capable of correctly labeling the consonant. Compared to baseline, our models were able to accurately estimate the SNR$_{90}$~intelligibility measure with less than 2 [dB$^2$] Mean Squared Error (MSE) on average, while the baseline ASR-defined measure computes SNR$_{90}$~with a variance of 5.2 to 26.6 [dB$^2$], depending on the consonant.
