GlotNet-A Raw Waveform Model for the Glottal Excitation in Statistical Parametric Speech Synthesis

Research output: Contribution to journalArticle

Standard

Harvard

APA

Vancouver

Author

Bibtex - Download

@article{df20945ef1e84959a017f1b9a0d6dbee,
title = "GlotNet-A Raw Waveform Model for the Glottal Excitation in Statistical Parametric Speech Synthesis",
abstract = "Recently, generative neural network models which operate directly on raw audio, such as WaveNet, have improved the state of the art in text-to-speech synthesis (TTS). Moreover, there is increasing interest in using these models as statistical vocoders for generating speech waveforms from various acoustic features. However, there is also a need to reduce the model complexity, without compromising the synthesis quality. Previously, glottal pulseforms (i.e., time-domain waveforms corresponding to the source of human voice production mechanism) have been successfully synthesized in TTS by glottal vocoders using straightforward deep feedforward neural networks. Therefore, it is natural to extend the glottal waveform modeling domain to use the more powerful WaveNet-like architecture. Furthermore, due to their inherent simplicity, glottal excitation waveforms permit scaling down the waveform generator architecture. In this study, we present a raw waveform glottal excitation model, called GlotNet, and compare its performance with the corresponding direct speech waveform model, WaveNet, using equivalent architectures. The models are evaluated as part of a statistical parametric TTS system. Listening test results show that both approaches are rated highly in voice similarity to the target speaker, and obtain similar quality ratings with large models. Furthermore, when the model size is reduced, the quality degradation is less severe for GlotNet.",
keywords = "Acoustics, Vocoders, Speech synthesis, Computational modeling, Hidden Markov models, Neural networks, Glottal source model, text-to-speech, WaveNet",
author = "Lauri Juvela and Bajibabu Bollepalli and Vassilis Tsiaras and Paavo Alku",
year = "2019",
month = "6",
day = "1",
doi = "10.1109/TASLP.2019.2906484",
language = "English",
volume = "27",
pages = "1019--1030",
journal = "IEEE/ACM Transactions on Audio, Speech, and Language Processing",
issn = "2329-9290",
publisher = "IEEE Advancing Technology for Humanity",
number = "6",

}

RIS - Download

TY - JOUR

T1 - GlotNet-A Raw Waveform Model for the Glottal Excitation in Statistical Parametric Speech Synthesis

AU - Juvela, Lauri

AU - Bollepalli, Bajibabu

AU - Tsiaras, Vassilis

AU - Alku, Paavo

PY - 2019/6/1

Y1 - 2019/6/1

N2 - Recently, generative neural network models which operate directly on raw audio, such as WaveNet, have improved the state of the art in text-to-speech synthesis (TTS). Moreover, there is increasing interest in using these models as statistical vocoders for generating speech waveforms from various acoustic features. However, there is also a need to reduce the model complexity, without compromising the synthesis quality. Previously, glottal pulseforms (i.e., time-domain waveforms corresponding to the source of human voice production mechanism) have been successfully synthesized in TTS by glottal vocoders using straightforward deep feedforward neural networks. Therefore, it is natural to extend the glottal waveform modeling domain to use the more powerful WaveNet-like architecture. Furthermore, due to their inherent simplicity, glottal excitation waveforms permit scaling down the waveform generator architecture. In this study, we present a raw waveform glottal excitation model, called GlotNet, and compare its performance with the corresponding direct speech waveform model, WaveNet, using equivalent architectures. The models are evaluated as part of a statistical parametric TTS system. Listening test results show that both approaches are rated highly in voice similarity to the target speaker, and obtain similar quality ratings with large models. Furthermore, when the model size is reduced, the quality degradation is less severe for GlotNet.

AB - Recently, generative neural network models which operate directly on raw audio, such as WaveNet, have improved the state of the art in text-to-speech synthesis (TTS). Moreover, there is increasing interest in using these models as statistical vocoders for generating speech waveforms from various acoustic features. However, there is also a need to reduce the model complexity, without compromising the synthesis quality. Previously, glottal pulseforms (i.e., time-domain waveforms corresponding to the source of human voice production mechanism) have been successfully synthesized in TTS by glottal vocoders using straightforward deep feedforward neural networks. Therefore, it is natural to extend the glottal waveform modeling domain to use the more powerful WaveNet-like architecture. Furthermore, due to their inherent simplicity, glottal excitation waveforms permit scaling down the waveform generator architecture. In this study, we present a raw waveform glottal excitation model, called GlotNet, and compare its performance with the corresponding direct speech waveform model, WaveNet, using equivalent architectures. The models are evaluated as part of a statistical parametric TTS system. Listening test results show that both approaches are rated highly in voice similarity to the target speaker, and obtain similar quality ratings with large models. Furthermore, when the model size is reduced, the quality degradation is less severe for GlotNet.

KW - Acoustics

KW - Vocoders

KW - Speech synthesis

KW - Computational modeling

KW - Hidden Markov models

KW - Neural networks

KW - Glottal source model

KW - text-to-speech

KW - WaveNet

UR - http://www.scopus.com/inward/record.url?scp=85064621868&partnerID=8YFLogxK

U2 - 10.1109/TASLP.2019.2906484

DO - 10.1109/TASLP.2019.2906484

M3 - Article

VL - 27

SP - 1019

EP - 1030

JO - IEEE/ACM Transactions on Audio, Speech, and Language Processing

JF - IEEE/ACM Transactions on Audio, Speech, and Language Processing

SN - 2329-9290

IS - 6

M1 - 8675543

ER -

ID: 32741487