Robust MelGAN: A robust universal neural vocoder for high-fidelity TTS

TL;DR

Robust MelGAN is a universal neural vocoder designed for high-fidelity TTS, employing novel dropout and data augmentation techniques to enhance robustness and generalization across diverse data sources.

Contribution

The paper introduces a robust version of MelGAN with specialized dropout and data augmentation methods to improve universal applicability and sound quality in TTS systems.

Findings

Significantly improves sound quality across various data types.

Reduces metallic sound issues in neural vocoders.

Maintains speaker similarity with enhanced robustness.

Abstract

In current two-stage neural text-to-speech (TTS) paradigm, it is ideal to have a universal neural vocoder, once trained, which is robust to imperfect mel-spectrogram predicted from the acoustic model. To this end, we propose Robust MelGAN vocoder by solving the original multi-band MelGAN's metallic sound problem and increasing its generalization ability. Specifically, we introduce a fine-grained network dropout strategy to the generator. With a specifically designed over-smooth handler which separates speech signal intro periodic and aperiodic components, we only perform network dropout to the aperodic components, which alleviates metallic sounding and maintains good speaker similarity. To further improve generalization ability, we introduce several data augmentation methods to augment fake data in the discriminator, including harmonic shift, harmonic noise and phase noise. Experiments show that Robust MelGAN can be used as a universal vocoder, significantly improving sound quality in TTS systems built on various types of data.