FloWaveNet : A Generative Flow for Raw Audio

TL;DR

FloWaveNet is a flow-based generative model for raw audio that enables real-time synthesis with a simple, single-stage training process, achieving high-quality sound comparable to more complex models.

Contribution

It introduces FloWaveNet, a novel flow-based model that simplifies training and inference for raw audio synthesis without auxiliary losses or two-stage training.

Findings

Real-time raw audio synthesis achieved with FloWaveNet

Single-stage training with maximum likelihood loss

Comparable audio quality to two-stage models

Abstract

Most modern text-to-speech architectures use a WaveNet vocoder for synthesizing high-fidelity waveform audio, but there have been limitations, such as high inference time, in its practical application due to its ancestral sampling scheme. The recently suggested Parallel WaveNet and ClariNet have achieved real-time audio synthesis capability by incorporating inverse autoregressive flow for parallel sampling. However, these approaches require a two-stage training pipeline with a well-trained teacher network and can only produce natural sound by using probability distillation along with auxiliary loss terms. We propose FloWaveNet, a flow-based generative model for raw audio synthesis. FloWaveNet requires only a single-stage training procedure and a single maximum likelihood loss, without any additional auxiliary terms, and it is inherently parallel due to the characteristics of generative flow. The model can efficiently sample raw audio in real-time, with clarity comparable to previous two-stage parallel models. The code and samples for all models, including our FloWaveNet, are publicly available.