Computer Science > Sound
[Submitted on 10 Apr 2021 (v1), last revised 27 Jun 2021 (this version, v3)]
Title:Unified Source-Filter GAN: Unified Source-filter Network Based On Factorization of Quasi-Periodic Parallel WaveGAN
View PDFAbstract:We propose a unified approach to data-driven source-filter modeling using a single neural network for developing a neural vocoder capable of generating high-quality synthetic speech waveforms while retaining flexibility of the source-filter model to control their voice characteristics. Our proposed network called unified source-filter generative adversarial networks (uSFGAN) is developed by factorizing quasi-periodic parallel WaveGAN (QPPWG), one of the neural vocoders based on a single neural network, into a source excitation generation network and a vocal tract resonance filtering network by additionally implementing a regularization loss. Moreover, inspired by neural source filter (NSF), only a sinusoidal waveform is additionally used as the simplest clue to generate a periodic source excitation waveform while minimizing the effect of approximations in the source filter model. The experimental results demonstrate that uSFGAN outperforms conventional neural vocoders, such as QPPWG and NSF in both speech quality and pitch controllability.
Submission history
From: Reo Yoneyama [view email][v1] Sat, 10 Apr 2021 02:38:26 UTC (1,778 KB)
[v2] Tue, 13 Apr 2021 03:14:06 UTC (1,778 KB)
[v3] Sun, 27 Jun 2021 11:30:56 UTC (1,786 KB)
Current browse context:
cs.SD
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.