End-to-End Multi-Task Denoising for the Joint Optimization of Perceptual Speech Metrics
This work addresses speech enhancement for noisy audio processing, offering incremental improvements by combining known techniques to better align with perceptual metrics.
The paper tackled the problems of spectrum and metric mismatches in speech enhancement by proposing an end-to-end denoising framework that optimizes time-domain signals and uses perceptual loss functions, resulting in significant improvements in SDR, PESQ, and STOI over existing methods.
Although supervised learning based on a deep neural network has recently achieved substantial improvement on speech enhancement, the existing schemes have either of two critical issues: spectrum or metric mismatches. The spectrum mismatch is a well known issue that any spectrum modification after short-time Fourier transform (STFT), in general, cannot be fully recovered after inverse short-time Fourier transform (ISTFT). The metric mismatch is that a conventional mean square error (MSE) loss function is typically sub-optimal to maximize perceptual speech measure such as signal-to-distortion ratio (SDR), perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI). This paper presents a new end-to-end denoising framework. First, the network optimization is performed on the time-domain signals after ISTFT to avoid the spectrum mismatch. Second, three loss functions based on SDR, PESQ and STOI are proposed to minimize the metric mismatch. The experimental result showed the proposed denoising scheme significantly improved SDR, PESQ and STOI performance over the existing methods. Moreover, the proposed scheme also provided good generalization performance over generative denoising models on the perceptual speech metrics not used as a loss function during training.