Statistical Parametric Speech Synthesis Using Generative Adversarial Networks Under A Multi-task Learning Framework
This work addresses speech synthesis quality for applications like text-to-speech systems, but it is incremental as it builds on existing GAN and multi-task learning methods.
The paper tackled the problem of improving synthesized speech naturalness in statistical parametric speech synthesis by proposing a multi-task learning framework that combines traditional acoustic loss with a GAN's discriminative loss, resulting in more natural speech that better satisfies human perception as shown in listening tests.
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.