On Loss Functions for Supervised Monaural Time-Domain Speech Enhancement
Top Cited Papers
- 23 January 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/ACM Transactions on Audio, Speech, and Language Processing
- Vol. 28 (23299290) , 825-838
Abstract
Many deep learning-based speech enhancement algorithms are designed to minimize the mean-square error (MSE) in some transform domain between a predicted and a target speech signal. However, optimizing for MSE does not necessarily guarantee high speech quality or intelligibility, which is the ultimate goal of many speech enhancement algorithms. Additionally, only little is known about the impact of the loss function on the emerging class of time-domain deep learning-based speech enhancement systems. We study how popular loss functions influence the performance of time-domain deep learning-based speech enhancement systems. First, we demonstrate that perceptually inspired loss functions might be advantageous over classical loss functions like MSE. Furthermore, we show that the learning rate is a crucial design parameter even for adaptive gradient-based optimizers, which has been generally overlooked in the literature. Also, we found that waveform matching performance metrics must be used with caution as they in certain situations can fail completely. Finally, we show that a loss function based on scale-invariant signal-to-distortion ratio (SI-SDR) achieves good general performance across a range of popular speech enhancement evaluation metrics, which suggests that SI-SDR is a good candidate as a general-purpose loss function for speech enhancement systems.Keywords
All Related Versions
This publication has 54 references indexed in Scilit:
- Large-scale training to increase speech intelligibility for hearing-impaired listeners in novel noisesThe Journal of the Acoustical Society of America, 2016
- Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification2015
- The third ‘CHiME’ speech separation and recognition challenge: Dataset, task and baselines2015
- An algorithm to increase speech intelligibility for hearing-impaired listeners in novel segments of the same noise typeThe Journal of the Acoustical Society of America, 2015
- On Training Targets for Supervised Speech SeparationIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2014
- DFT-Domain Based Single-Microphone Noise Reduction for Speech Enhancement: A Survey of the State of the ArtSynthesis Lectures on Speech and Audio Processing, 2013
- A classification based approach to speech segregationThe Journal of the Acoustical Society of America, 2012
- Preferred signal path delay and high-pass cut-off in open fittingsInternational Journal of Audiology, 2010
- An algorithm that improves speech intelligibility in noise for normal-hearing listenersThe Journal of the Acoustical Society of America, 2009
- Speech enhancement using a minimum-mean square error short-time spectral amplitude estimatorIEEE Transactions on Acoustics, Speech, and Signal Processing, 1984