Improving the expressiveness of neural vocoding with non-affine Normalizing Flows

TL;DR

This paper enhances neural vocoding by replacing affine transformations with non-affine invertible functions in Normalizing Flows, significantly improving speech naturalness and quality across multiple styles and languages.

Contribution

It introduces a non-affine transformation extension to Parallel Wavenet, boosting expressive capacity and perceptual quality in neural vocoding beyond existing affine-based methods.

Findings

Achieves 10% closer naturalness to original recordings.

Reduces spectral distance and cross-entropy by over 6%.

Outperforms state-of-the-art neural vocoders significantly.

Abstract

This paper proposes a general enhancement to the Normalizing Flows (NF) used in neural vocoding. As a case study, we improve expressive speech vocoding with a revamped Parallel Wavenet (PW). Specifically, we propose to extend the affine transformation of PW to the more expressive invertible non-affine function. The greater expressiveness of the improved PW leads to better-perceived signal quality and naturalness in the waveform reconstruction and text-to-speech (TTS) tasks. We evaluate the model across different speaking styles on a multi-speaker, multi-lingual dataset. In the waveform reconstruction task, the proposed model closes the naturalness and signal quality gap from the original PW to recordings by $10\%$, and from other state-of-the-art neural vocoding systems by more than $60\%$. We also demonstrate improvements in objective metrics on the evaluation test set with L2 Spectral Distance and Cross-Entropy reduced by $3\%$ and $6\unicode{x2030}$ comparing to the affine PW. Furthermore, we extend the probability density distillation procedure proposed by the original PW paper, so that it works with any non-affine invertible and differentiable function.