Logo image
Back
Voice Conversion for Persons with Amyotrophic Lateral Sclerosis
Journal article   Peer reviewed

Voice Conversion for Persons with Amyotrophic Lateral Sclerosis

Yunxin Zhao, Mili Kuruvilla-Dugdale and Minguang Song
IEEE journal of biomedical and health informatics, Vol.24(10), pp.2942-2949
10/01/2020
DOI: 10.1109/JBHI.2019.2961844
PMCID: 7314644
PMID: 31880571

View Online

Abstract

Amyotrophic lateral sclerosis (ALS) results in progressive paralysis of voluntary muscles throughout the body. As speech deteriorates, individuals rely on pre-programmed messages available on commercial speech generating devices to communicate using one of the generic electronic voices on the device. To replace these generic voices and restore vocal identity, our aim is to develop personalized voices for people with ALS via the approach of voice conversion. The task is challenging because very few people have large quantities of their premorbid healthy speech recorded. Therefore, we have to rely on small quantities of dysarthric speech concomitant with an individual's disease stage. Further, progressive fatigue prohibits acquisition of large speech datasets and individuals display a range of dysarthria severities resulting from breathing, voice, articulation, resonance, and prosody disturbances. As the first step to address these problems, we use healthy source speakers and propose the approach of combining a structured sparse spectral transform with multiple linear regression-based frequency warping prediction for spectral conversion, and interpolating the transformed spectral frames for speech rate modification. Our experimental data included four healthy source speakers from the ARCTIC dataset, and four target ALS speakers with mild to severe dysarthria, forming 16 speaker pairs. Subjective listening evaluations showed that on average, (i) the proposed approach improved speech intelligibility by about 80% over the target speakers' speech, (ii) the converted voice was 3 times more similar to the target speakers' speech than to the source speakers' speech, and (iii) the converted speech quality was close to the MOS scale "good" relative to the source speakers' speech being "excellent."
 Computer Science Computer Science, Information Systems Computer Science, Interdisciplinary Applications Life Sciences & Biomedicine Mathematical & Computational Biology Medical Informatics Science & Technology Technology

Details

Metrics

66 Record Views
16 Times Cited - Web of Science
Logo image