Effect Of Noise Suppression Losses On Speech Distortion And ASR Performance

Abstract

Deep learning based speech enhancement has made rapid development towards improving quality, while models are becoming more compact and usable for real-time on-the-edge inference. However, the speech quality scales directly with the model size, and small models are often still unable to achieve sufficient quality. Furthermore, the introduced speech distortion and artifacts greatly harm speech quality and intelligibility, and often significantly degrade automatic speech recognition (ASR) rates. In this work, we shed light on the success of the spectral complex compressed mean squared error (MSE) loss, and how its magnitude and phase-aware terms are related to the speech distortion vs. noise reduction trade off. We further investigate integrating pre-trained reference-less predictors for mean opinion score (MOS) and word error rate (WER), and pre-trained embeddings on ASR and sound event detection. Our analyses reveal that none of the pre-trained networks added significant performance ov

Stats

• citations26
• S2 citations—
• github stars0
• HF likes0
• heat score10.74
• arxiv keybraun2021effect

Related papers

• A Consolidated View Of Loss Functions For Supervised Deep Learning-based Speech Enhancement (2020)13.93
• Cheapnet: Improving Light-weight Speech Enhancement Network By Projected Loss Function (2023)0.00
• Bridging The Gap Between Monaural Speech Enhancement And Recognition With Distortion-independent Acoustic Modeling (2019)7.50
• Weighted Speech Distortion Losses For Neural-network-based Real-time Speech Enhancement (2020)14.51
• A Modulation-domain Loss For Neural-network-based Real-time Speech Enhancement (2021)8.09
• Rethinking Processing Distortions: Disentangling The Impact Of Speech Enhancement Errors On Speech Recognition Performance (2024)8.35
• Speech And Noise Dual-stream Spectrogram Refine Network With Speech Distortion Loss For Robust Speech Recognition (2023)5.24
• A Comparative Evaluation Of Deep Learning Models For Speech Enhancement In Real-world Noisy Environments (2025)0.00