Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain
This work addresses speech enhancement for audio processing applications, presenting an incremental improvement over prior generative diffusion models.
The authors tackled speech enhancement by extending score-based generative models to the complex STFT domain, proposing a novel training task that avoids noise distribution assumptions and results in improved performance.
Score-based generative models (SGMs) have recently shown impressive results for difficult generative tasks such as the unconditional and conditional generation of natural images and audio signals. In this work, we extend these models to the complex short-time Fourier transform (STFT) domain, proposing a novel training task for speech enhancement using a complex-valued deep neural network. We derive this training task within the formalism of stochastic differential equations (SDEs), thereby enabling the use of predictor-corrector samplers. We provide alternative formulations inspired by previous publications on using generative diffusion models for speech enhancement, avoiding the need for any prior assumptions on the noise distribution and making the training task purely generative which, as we show, results in improved enhancement performance.