Vocal effort compensation for MFCC feature extraction in a shouted versus normal speaker recognition task. (January 2019)

• Record Type: Journal Article
• Authors: Jokinen, Emma; Saeidi, Rahim; Kinnunen, Tomi; Alku, Paavo
• Is Part Of: Computer speech & language. Volume 53(2019)
• Page Start: 1
• Page End: 11
• Publication Date: 2019-01
• Subjects: Speaker recognition -- Vocal effort mismatch -- Shouted speech; Speech processing systems -- Periodicals; Automatic speech recognition -- Periodicals; Computers -- Periodicals; Linguistics -- Periodicals; Speech-Language Pathology -- Periodicals; Traitement automatique de la parole -- Périodiques; Reconnaissance automatique de la parole -- Périodiques; Automatic speech recognition; Speech processing systems; Electronic journals; Periodicals; 006.454
• DOI: 10.1016/j.csl.2018.06.002
• Languages: English
• ISSNs: 0885-2308

Abstract

Highlights: Speaker recognition from shouts is poor with systems trained using normal speech. Recognition drops due to vocal effort mismatch between training and testing. Two methods are proposed to compensate for vocal effort mismatch in speaker recognition. The techniques modify the all-pole power spectrum of the MFCC computation chain. The proposed methods improve identification rates in vocal effort mismatch. Abstract: In shouting, speakers use increased vocal effort to convey spoken messages over distance or above environmental noise. For automatic speaker recognition systems trained using normal speech, shouting causes a severe vocal effort mismatch between the enrollment and test hence reducing the recognition performance. In this study, two compensation methods are proposed to tackle the mismatch in a shouted versus normal speaker recognition task. These techniques are applied in the feature extraction stage of a speaker recognition system to modify the spectral envelopes of shouts to be closer to those in normal speech. The techniques modify the all-pole power spectrum of the MFCC computation chain with shouted-to-normal compensation filtering that is obtained using a GMM-based statistical mapping. In an evaluation using the state-of-the-art i-vector based recognition system, the proposed techniques provided considerable improvements in identification rates compared to the case when shouted speech spectra were not processed.