DiffPhase: Generative Diffusion-based STFT Phase Retrieval

TL;DR

This paper introduces DiffPhase, a diffusion probabilistic model tailored for STFT phase retrieval in speech processing, demonstrating superior performance over traditional methods in speech quality and intelligibility.

Contribution

The work adapts a diffusion-based speech enhancement model specifically for phase retrieval, showcasing its effectiveness in generating missing phase information.

Findings

DiffPhase outperforms classical phase retrieval methods.

Diffusion models are effective for imputation in phase retrieval.

Enhanced speech quality and intelligibility metrics.

Abstract

Diffusion probabilistic models have been recently used in a variety of tasks, including speech enhancement and synthesis. As a generative approach, diffusion models have been shown to be especially suitable for imputation problems, where missing data is generated based on existing data. Phase retrieval is inherently an imputation problem, where phase information has to be generated based on the given magnitude. In this work we build upon previous work in the speech domain, adapting a speech enhancement diffusion model specifically for STFT phase retrieval. Evaluation using speech quality and intelligibility metrics shows the diffusion approach is well-suited to the phase retrieval task, with performance surpassing both classical and modern methods.