Print Email Facebook Twitter An evaluation of objective measures for intelligibility prediction of time-frequency weighted noisy speech Title An evaluation of objective measures for intelligibility prediction of time-frequency weighted noisy speech Author Taal, C.H. Hendriks, R.C. Heusdens, R. Jensen, J. Faculty Electrical Engineering, Mathematics and Computer Science Department Mediamatics Date 2011-11-30 Abstract Existing objective speech-intelligibility measures are suitable for several types of degradation, however, it turns out that they are less appropriate in cases where noisy speech is processed by a time-frequency weighting. To this end, an extensive evaluation is presented of objective measure for intelligibility prediction of noisy speech processed with a technique called ideal time frequency (TF) segregation. In total 17 measures are evaluated, including four advanced speech-intelligibility measures (CSII, CSTI, NSEC, DAU), the advanced speech-quality measure (PESQ), and several frame-based measures (e.g., SSNR). Furthermore, several additional measures are proposed. The study comprised a total number of 168 different TF-weightings, including unprocessed noisy speech. Out of all measures, the proposed frame-based measure MCC gave the best results (??=?0.93). An additional experiment shows that the good performing measures in this study also show high correlation with the intelligibility of single-channel noise reduced speech. Subject acoustic noisespeech intelligibility To reference this document use: http://resolver.tudelft.nl/uuid:ff6fdb5c-7706-4dd4-90f3-60fa7f3f53d9 DOI https://doi.org/10.1121/1.3641373 Publisher Acoustical Society of America Embargo date 2012-05-30 ISSN 0001-4966 Source Journal of the Acoustical Society of America, 130 (5), 2011 Part of collection Institutional Repository Document type journal article Rights © 2011 The Author(s)Acoustical Society of America Files PDF Taal.pdf 999.8 KB Close viewer /islandora/object/uuid:ff6fdb5c-7706-4dd4-90f3-60fa7f3f53d9/datastream/OBJ/view