Phase Estimation from Noisy Observation

We recently proposed several phase estimation methods to estimate the clean phase from an observed single-channel noisy speech signal. The so-obtained phase estimates can be used to derive a phase-sensitive amplitude estimator or the iterative phase-aware closed-loop speech enhancement.

For a recent overview on phase estimation from noisy observation we refer to "Ch. 3 in the book".

M. Pirolt, J. Stahl, P. Mowlaee, V. Vorobiov, S. Barysenka, A. Davydov, "Phase Estimation in Single-Channel Speech Enhancement Using Phase Invariance Constraints", ICASSP, pp. 5585-5589, 2017.

P. Mowlaee, J. Stahl, J. Kulmer, "Iterative Joint MAP Single-Channel Speech Enhancement Given Non-uniform Phase Prior", Speech Communication, vol. 86, pp. 85-96, Feb. 2017.[Audio]

J. Stahl, P. Mowlaee, J. Kulmer, "Phase-Processing For Voice Activity Detection: A Statistical Approach", in Proc. European Signal Processing Conference (EUSIPCO 2016), Budapest, Hungary, September 2016.

A. Maly, P. Mowlaee, "On The Importance of Harmonic Phase Modification for Improved Speech Signal Reconstruction", in Proc. ICASSP 2016, pp. 584-588, Shanghai, China, March 2016.[Audio][Poster]

P. Mowlaee, J. Kulmer, "Harmonic Phase Estimation in Single-Channel Speech Enhancement Using Phase Decomposition and SNR Information", IEEE Transactions on Audio, Speech and Language, vol. 23, no. 9, pp. 1521-1532, Sept, 2015.

P. Mowlaee, "Phase Estimation From Theory to Practice", Tutorial session at INTERSPEECH 2015, Dresden, Germany, 2015. link

P. Mowlaee, J. Kulmer, "Phase Estimation in Single-Channel Speech Enhancement: Limits-Potential", IEEE Transactions on Audio, Speech and Language, vol. 23, Issue 8, pp. 1283-1294, August, 2015.

J. Kulmer, P. Mowlaee, "Harmonic Phase Estimation in Single-Channel Speech Enhancement Using Von Mises Distribution and Prior SNR", in Proc. ICASSP 2015, pp. 5063-5067, Brisbane, Australia, April 2015 [Matlab].

J. Kulmer, P. Mowlaee, "Phase Estimation in Single Channel Speech Enhancement Using Phase Decomposition", IEEE Signal Processing Letters, vol.22, no.5, pp.598,602, May 2015 [Matlab].

J. Kulmer, P. Mowlaee, M. K. Watanabe, "A probabilistic approach for phase estimation in single-channel speech enhancement using von mises phase priors", IEEE International Workshop on Machine Learning for Signal Processing (MLSP), vol., no., pp.1,6, 21-24 Sept. 2014.

P. Mowlaee, R. Saeidi, "Time-Frequency Constraints for Phase Estimation in Single-Channel Speech Enhancement", The International Workshop on Acoustic Signal Enhancement (IWAENC), 2014. [Audio]

C. Chacon, P. Mowlaee, "Least Squares Phase Estimation of Mixed Signals", 15th Annual Conference of the International Speech Communication Association, Singapore, 2014.

P. Mowlaee, R. Saeidi, M. Watanabe, "Show & Tell: Iterative Refinement of Amplitude and Phase in Single-channel Speech Enhancement", INTERSPEECH 2014, pp. 2134-2135, 2014

P. Mowlaee, M. K. Watanabe, R. Saeidi, "Show and Tell: Phase-Aware Single-Channel Speech Enhancement", 14th Annual Conference of the International Speech Communication Association, Lyon, France, pp. 1872-1874, 2013.

M. Watanabe, P. Mowlaee, "Iterative sinusoidal-based partial phase reconstruction in single-channel source separation", INTERSPEECH 2013. [Audio] [Matlab].

P. Mowlaee, R. Martin, "On Phase Importance in Parameter Estimation for Single-channel Source Separation”, The International Workshop on Acoustic Signal Enhancement (IWAENC), 2012. [Audio] [Matlab].

P. Mowlaee, R. Saeidi, R. Martin, "Phase estimation for signal reconstruction in single-channel source separation”, in Proc. 13th Annual Conference of the International Speech Communication Association, Interspeech 2012. [Audio] [Matlab].

The following examples demonstrate the phase estimation results and its impact on speech enhancement performance.
The following audio examples demonstrate the performance of the phase estimation methods for different noise and SNR scenarios.

Babble Noise at SNR = 0 (dB):

Clean Speech Noisy Speech
STFTPI Proposed
Clean phase Consistent Wiener Filter