LVCNet: Efficient Condition-Dependent Modeling Network for Waveform Generation

Abstract

In this paper, we propose a novel conditional convolution network, named location-variable convolution, to model the dependencies of the waveform sequence. Different from the use of unified convolution kernels in WaveNet to capture the dependencies of arbitrary waveform, the location-variable convolution uses convolution kernels with different coefficients to perform convolution operations on different waveform intervals, where the coefficients of kernels is predicted according to conditioning acoustic features, such as Mel-spectrograms. Based on location-variable convolutions, we design LVCNet for waveform generation, and apply it in Parallel WaveGAN to design more efficient vocoder. Experiments on the LJSpeech dataset show that our proposed model achieves a four-fold increase in synthesis speed compared to the original Parallel WaveGAN without any degradation in sound quality, which verifies the effectiveness of location-variable convolutions.

Accepted to ICASSP 2021. arXiv admin note: text overlap with arXiv:2012.01684

References in corpus (8)

• Denoising Diffusion Probabilistic Models
• WaveNet: A Generative Model for Raw Audio
• MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
• Parallel WaveNet: Fast High-Fidelity Speech Synthesis
• DiffWave: A Versatile Diffusion Model for Audio Synthesis
• WaveFlow: A Compact Flow-based Model for Raw Audio
• Neural Speech Synthesis with Transformer Network
• MelGlow: Efficient Waveform Generative Network Based on Location-Variable Convolution

Cited by in corpus (3)

• A Survey on Neural Speech Synthesis
• End-to-End Zero-Shot Voice Conversion with Location-Variable Convolutions
• Adversarial Learning of Intermediate Acoustic Feature for End-to-End Lightweight Text-to-Speech