The unimportance of phase in speech enhancement

By Deliang Wang, Jae S. Lim. 1982

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Vocos: Closing the gap between time-domain and Fourier-based neural vocoders for high-quality audio synthesis (2023, background): “Furthermore, the literature does not provide a definitive answer regarding the perceptual importance of phase-related information in speech (Wang and Lim, 1982; Paliwal et al., 2011).”
SEGAN: Speech Enhancement Generative Adversarial Network (2017, background): “However, further studies [18] show that signiﬁcant improvements of speech quality are possible, especially when a clean phase spectrum is known.”
Complex Ratio Masking for Monaural Speech Separation (2015, background): “STOI measures objective intelligibility by computing the correlation of short-time temporal envelopes between clean and separated speech, resulting in scores in the range of [0] [1] where a higher score indicates better intelligibility .”
The importance of phase in speech enhancement (2011, methodology): “Thus employing clean phase spectrum provides little to no improvement in speech quality for AMS settings similar to those used in (Wang and Lim, 1982).”
New insights into the noise reduction Wiener filter (2006, methodology): “A variety of approaches have been developed, including Wiener filter [3], [14]–[19], spectral or cepstral restoration [17], [20]–[27], signal subspace [28]–[35], parametric-model-based method [36]–[38], and statistical-model-based method [5], [39]–[46].”
Phase Processing for Single-Channel Speech Enhancement: History and recent advances (2015, methodology): “To explore the relevance of phase estimation for speech enhancement, Wang and Lim [10] performed listening experiments where the magnitude of a noisy speech signal at a certain signal-to-noise ratio (SNR) was combined with the phase of the same speech signal but distorted by noise at a different SNR”
A two-stage algorithm for one-microphone reverberant speech enhancement (2006, background): “Many objective speech quality measures (for example, see [30]) are solely based on the magnitude spectrum of speech, in part motivated by the study of Wang and Lim [33] showing that phase distortion is not important for enhancing speech mixed with a moderate level of white noise.”
Learning Complex Spectral Mapping with Gated Convolutional Recurrent Networks for Monaural Speech Enhancement (2019, background): “Second, it was believed that phase enhancement is not important for speech enhancement [37].”
A Reverberation-Time-Aware Approach to Speech Dereverberation Based on Deep Neural Networks (2017, methodology): “The generalization capabilities of the proposed DNN models are illustrated in Section V. Finally we summarize our findings in Section VI.”
Speech enhancement for non-stationary noise environments (2001, background): “The phase of the noisy speech is not processed [27], since the optimal phase estimate is the noisy phase itself (in fact, the minimum mean-square error estimator of the complex exponential of the phase, under unity modulus constraint, is the complex exponential of the noisy phase [7]).”