Title :
A non-intrusive PESQ measure
Author :
Sharma, Dushyant ; Meredith, Lisa ; Lainez, Jose ; Barreda, Daniel ; Naylor, Patrick A.
Author_Institution :
Voicemail-To-Text Res., Nuance Commun. Inc., Marlow, UK
Abstract :
We present NISQ, a data-driven non-intrusive speech quality measure that has been trained to predict the PESQ score for a given speech signal. NISQ is based on feature extraction and a binary tree regression based model. A training method using the intrusive PESQ algorithm to automatically label large quantities of speech data is presented and utilized. Our method is shown to predict PESQ with an RMS error of 0.49 on our test database.
Keywords :
mean square error methods; regression analysis; signal processing; speech processing; trees (mathematics); NISQ; PESQ score; RMS error; binary tree regression based model; data-driven nonintrusive speech quality measure; feature extraction; intrusive PESQ algorithm; nonintrusive PESQ measure; speech data; speech signal; test database; training method; Databases; Quality assessment; Signal processing algorithms; Speech; Speech processing; Training; CART; Non-Intrusive; PESQ; Speech Quality;
Conference_Titel :
Signal and Information Processing (GlobalSIP), 2014 IEEE Global Conference on
Conference_Location :
Atlanta, GA
DOI :
10.1109/GlobalSIP.2014.7032266
