Improving Resource-Efficient Speech Enhancement via Neural Differentiable DSP Vocoder Refinement

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Guimarães, Heitor R., Tan, Ke, Azcarreta, Juan, Alvarez, Jesus, Agrawal, Prabhav, Pandey, Ashutosh, Xu, Buye
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908496396550144
author Guimarães, Heitor R.
Tan, Ke
Azcarreta, Juan
Alvarez, Jesus
Agrawal, Prabhav
Pandey, Ashutosh
Xu, Buye
author_facet Guimarães, Heitor R.
Tan, Ke
Azcarreta, Juan
Alvarez, Jesus
Agrawal, Prabhav
Pandey, Ashutosh
Xu, Buye
contents Deploying speech enhancement (SE) systems in wearable devices, such as smart glasses, is challenging due to the limited computational resources on the device. Although deep learning methods have achieved high-quality results, their computational cost limits their feasibility on embedded platforms. This work presents an efficient end-to-end SE framework that leverages a Differentiable Digital Signal Processing (DDSP) vocoder for high-quality speech synthesis. First, a compact neural network predicts enhanced acoustic features from noisy speech: spectral envelope, fundamental frequency (F0), and periodicity. These features are fed into the DDSP vocoder to synthesize the enhanced waveform. The system is trained end-to-end with STFT and adversarial losses, enabling direct optimization at the feature and waveform levels. Experimental results show that our method improves intelligibility and quality by 4% (STOI) and 19% (DNSMOS) over strong baselines without significantly increasing computation, making it well-suited for real-time applications.
format Preprint
id arxiv_https___arxiv_org_abs_2508_14709
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Improving Resource-Efficient Speech Enhancement via Neural Differentiable DSP Vocoder Refinement
Guimarães, Heitor R.
Tan, Ke
Azcarreta, Juan
Alvarez, Jesus
Agrawal, Prabhav
Pandey, Ashutosh
Xu, Buye
Audio and Speech Processing
Sound
Deploying speech enhancement (SE) systems in wearable devices, such as smart glasses, is challenging due to the limited computational resources on the device. Although deep learning methods have achieved high-quality results, their computational cost limits their feasibility on embedded platforms. This work presents an efficient end-to-end SE framework that leverages a Differentiable Digital Signal Processing (DDSP) vocoder for high-quality speech synthesis. First, a compact neural network predicts enhanced acoustic features from noisy speech: spectral envelope, fundamental frequency (F0), and periodicity. These features are fed into the DDSP vocoder to synthesize the enhanced waveform. The system is trained end-to-end with STFT and adversarial losses, enabling direct optimization at the feature and waveform levels. Experimental results show that our method improves intelligibility and quality by 4% (STOI) and 19% (DNSMOS) over strong baselines without significantly increasing computation, making it well-suited for real-time applications.
title Improving Resource-Efficient Speech Enhancement via Neural Differentiable DSP Vocoder Refinement
topic Audio and Speech Processing
Sound
url https://arxiv.org/abs/2508.14709