Nebula: F0 Estimation and Voicing Detection by Modeling the Statistical Properties of Feature Extractors

TL;DR

This paper introduces a novel F0 and voicing detection method that models feature extractor behavior under noise using Gaussian mixture models trained on synthetic data, improving accuracy in speech analysis.

Contribution

It presents a new approach that models feature extractor behavior with GMMs trained on artificial data, avoiding direct speech modeling and enhancing robustness.

Findings

Lower gross error rates than state-of-the-art methods.

Effective training on synthetic data.

Improved F0 estimation accuracy.

Abstract

A F0 and voicing status estimation algorithm for high quality speech analysis/synthesis is proposed. This problem is approached from a different perspective that models the behavior of feature extractors under noise, instead of directly modeling speech signals. Under time-frequency locality assumptions, the joint distribution of extracted features and target F0 can be characterized by training a bank of Gaussian mixture models (GMM) on artificial data generated from Monte-Carlo simulations. The trained GMMs can then be used to generate a set of conditional distributions on the predicted F0, which are then combined and post-processed by Viterbi algorithm to give a final F0 trajectory. Evaluation on CSTR and CMU Arctic speech databases shows that the proposed method, trained on fully synthetic data, achieves lower gross error rates than state-of-the-art methods.