Generative adversarial network-based glottal waveform model for statistical parametric speech synthesis

 |  Login

Show simple item record

dc.contributor Aalto-yliopisto fi
dc.contributor Aalto University en
dc.contributor.author Bollepalli, Bajibabu
dc.contributor.author Juvela, Lauri
dc.contributor.author Alku, Paavo
dc.date.accessioned 2017-11-21T13:39:20Z
dc.date.available 2017-11-21T13:39:20Z
dc.date.issued 2017-08
dc.identifier.citation Bollepalli , B , Juvela , L & Alku , P 2017 , Generative adversarial network-based glottal waveform model for statistical parametric speech synthesis . in Proceedings of Interspeech 2017 . Interspeech: Annual Conference of the International Speech Communication Association , International Speech Communication Association , pp. 3394-3398 , Interspeech , Stockholm , Sweden , 20-24 August . DOI: 10.21437/Interspeech.2017-1288 en
dc.identifier.issn 1990-9772
dc.identifier.other PURE UUID: d46f19a6-879f-41b9-9142-032ba6e624ab
dc.identifier.other PURE ITEMURL: https://research.aalto.fi/en/publications/generative-adversarial-networkbased-glottal-waveform-model-for-statistical-parametric-speech-synthesis(d46f19a6-879f-41b9-9142-032ba6e624ab).html
dc.identifier.other PURE FILEURL: https://research.aalto.fi/files/15742197/bollepalli_interspeech1288.pdf
dc.identifier.uri https://aaltodoc.aalto.fi/handle/123456789/28868
dc.description.abstract Recent studies have shown that text-to-speech synthesis quality can be improved by using glottal vocoding. This refers to vocoders that parameterize speech into two parts, the glottal excitation and vocal tract, that occur in the human speech production apparatus. Current glottal vocoders generate the glottal excitation waveform by using deep neural networks (DNNs). However, the squared error-based training of the present glottal excitation models is limited to generating conditional average waveforms, which fails to capture the stochastic variation of the waveforms. As a result, shaped noise is added as post-processing. In this study, we propose a new method for predicting glottal waveforms by generative adversarial networks (GANs). GANs are generative models that aim to embed the data distribution in a latent space, enabling generation of new instances very similar to the original by randomly sampling the latent distribution. The glottal pulses generated by GANs show a stochastic component similar to natural glottal pulses. In our experiments, we compare synthetic speech generated using glottal waveforms produced by both DNNs and GANs. The results show that the newly proposed GANs achieve synthesis quality comparable to that of widely-used DNNs, without using an additive noise component. en
dc.format.extent 5
dc.format.extent 3394-3398
dc.format.mimetype application/pdf
dc.language.iso en en
dc.relation.ispartof Interspeech en
dc.relation.ispartofseries Proceedings of Interspeech 2017 en
dc.relation.ispartofseries Interspeech: Annual Conference of the International Speech Communication Association en
dc.rights openAccess en
dc.subject.other 213 Electronic, automation and communications engineering, electronics en
dc.title Generative adversarial network-based glottal waveform model for statistical parametric speech synthesis en
dc.type A4 Artikkeli konferenssijulkaisussa fi
dc.description.version Peer reviewed en
dc.contributor.department Department of Signal Processing and Acoustics
dc.subject.keyword Glottal souce modelling
dc.subject.keyword GAN
dc.subject.keyword TTS
dc.subject.keyword DNN
dc.subject.keyword 213 Electronic, automation and communications engineering, electronics
dc.identifier.urn URN:NBN:fi:aalto-201711217689
dc.identifier.doi 10.21437/Interspeech.2017-1288
dc.type.version publishedVersion


Files in this item

Files Size Format View

There are no open access files associated with this item.

This item appears in the following Collection(s)

Show simple item record

Search archive


Advanced Search

article-iconSubmit a publication

Browse