Statistical Parametric Speech Synthesis Using Generative Adversarial Networks Under A Multi-task Learning Framework
arXiv:1707.01670
Abstract
In this paper, we aim at improving the performance of synthesized speech in statistical parametric speech synthesis (SPSS) based on a generative adversarial network (GAN). In particular, we propose a novel architecture combining the traditional acoustic loss function and the GAN's discriminative loss under a multi-task learning (MTL) framework. The mean squared error (MSE) is usually used to estimate the parameters of deep neural networks, which only considers the numerical difference between the raw audio and the synthesized one. To mitigate this problem, we introduce the GAN as a second task to determine if the input is a natural speech with specific conditions. In this MTL framework, the MSE optimization improves the stability of GAN, and at the same time GAN produces samples with a distribution closer to natural speech. Listening tests show that the multi-task architecture can generate more natural speech that satisfies human perception than the conventional methods.
Submitted to Automatic Speech Recognition and Understanding (ASRU) 2017 Workshop
References in corpus (9)
- Adam: A Method for Stochastic Optimization
- TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
- Conditional Generative Adversarial Nets
- BEGAN: Boundary Equilibrium Generative Adversarial Networks
- MidiNet: A Convolutional Generative Adversarial Network for Symbolic-domain Music Generation
- Variational Approaches for Auto-Encoding Generative Adversarial Networks
- Deep Voice 2: Multi-Speaker Neural Text-to-Speech
- Voice Conversion from Unaligned Corpora using Variational Autoencoding Wasserstein Generative Adversarial Networks
- Tacotron: Towards End-to-End Speech Synthesis