A system for creating personalized synthetic voices

Yarrington, Debra; Pennington, Christopher; Gray, John; Bunnell, H. Timothy

doi:10.1145/1090785.1090827

Cited by 9 publications

(3 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A user must record around 1,800 utterances that fully cover a set of diphones (which is the second half of one phone plus the first half of the following phone) and can upload the recordings to the developers. Within a short time, a personalized TTS voice, based on diphone concatenation synthesis, which was originally developed in the 1980s, becomes available for use [7].…”

Section: Commercial Voice Banking Servicementioning

confidence: 99%

Speech synthesis technologies for individuals with vocal disabilities: Voice banking and reconstruction

Yamagishi

Veaux

King

et al. 2012

Acoust. Sci. & Tech.

View full text Add to dashboard Cite

Section: Commercial Voice Banking Servicementioning

confidence: 99%

Speech synthesis technologies for individuals with vocal disabilities: Voice banking and reconstruction

Yamagishi

Veaux

King

et al. 2012

Acoust. Sci. & Tech.

View full text Add to dashboard Cite

“…In order to build personalized VOCAs, several attempts have been made to capture the voice before it is lost, using a process known as voice banking. One example of this approach is ModelTalker [3], a free voice building service that can be used from any home computer in order to build a synthetic voice based on diphone concatenation, a technology developed in the 1980s. The user of this service has to record around 1800 utterances in order to fully cover the set of diphones and the naturalness of the synthetic speech is rather low.…”

Section: Introductionmentioning

confidence: 99%

A Comparison of Manual and Automatic Voice Repair for Individual with Vocal Disabilities

Veaux¹,

Yamagishi²,

King

2015

Proceedings of SLPAT 2015: 6th Workshop on Speech and Language Processing for Assistive Technologies

View full text Add to dashboard Cite

When individuals lose the ability to produce their own speech, due to degenerative diseases such as motor neurone disease (MND) or Parkinson's, they lose not only a functional means of communication but also a display of their individual and group identity. In order to build personalized synthetic voices, attempts have been made to capture the voice before it is lost, using a process known as voice banking. But, for some patients, the speech deterioration frequently coincides or quickly follows diagnosis. Using HMM-based speech synthesis, it is now possible to build personalized synthetic voices with minimal data recordings and even disordered speech. The power of this approach is that it is possible to use the patient's recordings to adapt existing voice models pre-trained on many speakers. When the speech has begun to deteriorate, the adapted voice model can be further modified in order to compensate for the disordered characteristics found in the patient's speech, we call this process "voice repair". In this paper we compare two methods of voice repair. The first method follows a trial and error approach and requires the expertise of a speech therapist. The second method is entirely automatic and based on some a priori statistical knowledge. A subjective evaluation shows that the automatic method achieves similar results than the manually controlled method.

show abstract

“…Most existing personalised VOCA devices (ModelTalker [2], Cereproc[3], Polluxstar, based on a hy-This research was supported by ESPRC Programme Grant, grant no. EP/I031022/1 (Natural Speech Technology) brid TTS [4] using both unit selection and statistical parametric speech synthesis [5]) are based on a voice banking approach which is the process of capturing the voice before it starts to degrade.…”

Section: Introductionmentioning

confidence: 99%

Reconstructing voices within the multiple-average-voice-model framework

Lanchantin¹,

Veaux²,

Gales³

et al. 2015

Interspeech 2015

View full text Add to dashboard Cite

Personalisation of voice output communication aids (VOCAs) allows to preserve the vocal identity of people suffering from speech disorders. This can be achieved by the adaptation of HMM-based speech synthesis systems using a small amount of adaptation data. When the voice has begun to deteriorate, reconstruction is still possible in the statistical domain by correcting the parameters of the models associated with the speech disorder. This can be done by substituting those with parameters from a donor's voice, at risk of losing part of the identity of the patient. Recently, the Multiple-Average-Voice-Model (Multiple AVM) framework has been proposed for speaker adaptation. Adaptation is performed via interpolation into a speaker eigenspace spanned by the mean vectors of speaker-adapted AVMs which can be tuned to the individual speaker. In this paper, we present the benefits of this framework for voice reconstruction: it requires only a very small amount of adaptation data, interpolation can be performed in a clean speech eigenspace and the resulting voice can be easily fine-tuned by acting on the interpolation weights. We illustrate our points with a subjective assessment of the reconstructed voice.

show abstract

A system for creating personalized synthetic voices

Cited by 9 publications

References 1 publication

Speech synthesis technologies for individuals with vocal disabilities: Voice banking and reconstruction

Speech synthesis technologies for individuals with vocal disabilities: Voice banking and reconstruction

A Comparison of Manual and Automatic Voice Repair for Individual with Vocal Disabilities

Reconstructing voices within the multiple-average-voice-model framework

Contact Info

Product

Resources

About