# Parametric Resynthesis with neural vocoders

## TL;DR

This paper introduces a noise suppression method that leverages neural vocoders to generate high-quality clean speech from noisy inputs, outperforming traditional source separation models in quality.

## Contribution

The study demonstrates the effectiveness of neural vocoders like WaveNet and WaveGlow for noise suppression, showing significant quality improvements over existing models.

## Key findings

- WaveNet and WaveGlow outperform Chimera++ in quality scores.
- Both neural vocoders surpass the oracle Wiener mask in subjective quality.
- WaveNet achieves the best subjective quality but is slower.

## Abstract

Noise suppression systems generally produce output speech with compromised quality. We propose to utilize the high quality speech generation capability of neural vocoders for noise suppression. We use a neural network to predict clean mel-spectrogram features from noisy speech and then compare two neural vocoders, WaveNet and WaveGlow, for synthesizing clean speech from the predicted mel spectrogram. Both WaveNet and WaveGlow achieve better subjective and objective quality scores than the source separation model Chimera++. Further, WaveNet and WaveGlow also achieve significantly better subjective quality ratings than the oracle Wiener mask. Moreover, we observe that between WaveNet and WaveGlow, WaveNet achieves the best subjective quality scores, although at the cost of much slower waveform generation.