Surgery of Speech Synthesis Models to Overcome the Scarcity of Training Data

Pierard, Arnaud; Erro, Daniel; Hernáez, Inmaculada; Navas, Eva; Dutoit, Thierry

doi:10.1007/978-3-319-49169-1_8

Cited by 2 publications

(2 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In general, it is required that the person performs the recordings before the first symptoms of the disease are noticeable in the voice, because as commented in the introduction, this can condition the final quality of the synthetic voice. If this has already happened, some providers also offer the possibility to repair the synthetic voice, applying model surgery techniques [36][37][38]. On the other hand, the cited systems also differ in the number of sentences needed to be recorded (from a few hundred to several thousands) as well as on the operating systems, browsers or applications where the final user will be able to install it.…”

Section: Voice Banking and Personalized Synthetic Voicesmentioning

confidence: 99%

Automatic Classification of Synthetic Voices for Voice Banking Using Objective Measures

et al. 2022

View full text Add to dashboard Cite

Speech is the most common way of communication among humans. People who cannot communicate through speech due to partial of total loss of the voice can benefit from Alternative and Augmentative Communication devices and Text to Speech technology. One problem of using these technologies is that the included synthetic voices might be impersonal and badly adapted to the user in terms of age, accent or even gender. In this context, the use of synthetic voices from voice banking systems is an attractive alternative. New voices can be obtained applying adaptation techniques using recordings from people with healthy voice (donors) or from the user himself/herself before losing his/her own voice. In this way, the goal is to offer a wide voice catalog to potential users. However, as there is no control over the recording or the adaptation processes, some method to control the final quality of the voice is needed. We present the work developed to automatically select the best synthetic voices using a set of objective measures and a subjective Mean Opinion Score evaluation. A prediction algorithm of the MOS has been build which correlates similarly to the most correlated individual measure.

show abstract

Section: Voice Banking and Personalized Synthetic Voicesmentioning

confidence: 99%

Automatic Classification of Synthetic Voices for Voice Banking Using Objective Measures

et al. 2022

View full text Add to dashboard Cite

show abstract

“…We also tried to palliate the pronunciation problems applying the adaptation techniques using only vowels described in [8][9], but the new synthetic voice, although with an improved pronunciation lost the personality of the speaker. We are also experimenting with model surgery on some phones [10], but the improvements are subtle.…”

Section: Personalisation For Dysarthric Voicesmentioning

confidence: 99%

RESTORE Project: REpair, STOrage and REhabilitation of speech

Hernáez¹,

Navas²,

Martín³

et al. 2018

IberSPEECH 2018

View full text Add to dashboard Cite

RESTORE is a project aimed to improve the quality of communication for people with difficulties producing speech, providing them with tools and alternative communication services. At the same time, progress will be made at the research of techniques for restoration and rehabilitation of disordered speech. The ultimate goal of the project is to offer new possibilities in the rehabilitation and reintegration into society of patients with speech pathologies, especially those laryngectomised, by designing new intervention strategies aimed to favour their communication with the environment and ultimately increase their quality of life.

show abstract

Surgery of Speech Synthesis Models to Overcome the Scarcity of Training Data

Cited by 2 publications

References 18 publications

Automatic Classification of Synthetic Voices for Voice Banking Using Objective Measures

Automatic Classification of Synthetic Voices for Voice Banking Using Objective Measures

RESTORE Project: REpair, STOrage and REhabilitation of speech

Contact Info

Product

Resources

About