Speech intelligibility prediction method using machine learning for outdoor public address systems

• Published in: The Journal of the Acoustical Society of America Vol. 140; no. 4; p. 3192
• Main Authors: Kobayashi, Yosuke, Ohta, Kengo, Kondo, Kazuhiro, Sakamoto, Shuichi
• Format: Journal Article
• Language: English
• Published: 01.10.2016
• ISSN: 0001-4966; 1520-8524
• DOI: 10.1121/1.4970044

Subjective speech intelligibility assessment is important for the development of outdoor public address system. However, as this assessment is difficult in many cases, we propose an objective speech intelligibility evaluation system that includes a machine learning technique. In this talk, we have proved a subjective evaluation and objective prediction of speech intelligibility using the outdoor public address systems at 10 locations in Sendai City, where impulse responses were recorded after the Great East Japan Earthquake. First, the results of the subjective intelligibility evaluation by different test word lists with the same sound field conditions showed that the root mean squared error (RMSE) was very small, not exceeding 7.0%. Next, we generated the intelligibility prediction model trained with true/false results of 22 subjects using the support vector machine (SVM). This prediction model extracted the feature vector, using the ITU-T P.563 speech quality feature set of the test speech signal. We evaluated the predictive performance of the prediction model using data that was not used in training, and the RMSE obtained was 4.0%. This result was shown to be highly accurate with a value even less than the subject experiment result.