A signal subspace approach for speech enhancement

By Yariv Ephraim, H. V. Trees. 1993

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

SEGAN: Speech Enhancement Generative Adversarial Network (2017, methodology): “Neural networks have been also applied to speech enhancement since the 80s [11, 12].”
Speech Enhancement in Multiple-Noise Conditions Using Deep Neural Networks (2016, neutral): “Other popular classical methods include signal-subspace based methods [7] [8].”
A Supervised Learning Approach to Monaural Segregation of Reverberant Speech (2009, methodology): “Various methods have been proposed for monaural speech enhancement or separation, including spectral subtraction [8], Wiener ﬁltering [13,33], minimum mean square error (MMSE) estimation [18], and subspace analysis [19].”
Monaural speech segregation based on pitch tracking and amplitude modulation (2004, methodology): “Another way to deal with interference is to perform eigen-decomposition on an acoustic mixture and then apply subspace analysis to remove part of the interference [13].”
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].”
Multichannel Eigenspace Beamforming in a Reverberant Noisy Environment With Multiple Interfering Speech Signals (2009, background): “Ephraim and Van Trees [22] considered the single-microphone scenario.”
A noise-estimation algorithm for highly non-stationary environments (2006, methodology): “…as it is needed, for instance, to evaluate the Wiener ﬁlter in the Wiener algorithms (Lim and Oppenheim, 1978) or to estimate the a priori SNR in the MMSE algorithms (Ephraim and Malah, 1984) or to estimate the noise covariance matrix in the subspace algorithms (Ephraim and Van Trees, 1993).”
On the Importance of the Pearson Correlation Coefficient in Noise Reduction (2008, background): “Let us choose , we have SNR (37) (38) (39) From this filter, we can deduce another interesting property of the SPCC. Property 10: We have SNR SNR (40) Since SNR SNR SNR, this implies that (41) which means that the filter yields less distortion to the clean speech signal than to the noise signal .”
GSVD-based optimal filtering for single and multimicrophone speech enhancement (2002, methodology): “In [11], it has already been noted that these negative values will always be obtained when an unbiased nonperfect estimator is used.”
The importance of phase in speech enhancement (2011, methodology): “These can be grouped into spectral subtraction (Boll, 1979; Berouti et al., 1979), minimum mean-square error (MMSE) estimation (Ephraim and Malah, 1984, 1985), Wiener filtering (linear MMSE) (Wiener , 1949), Kalman filtering (Paliwal and Basu, 1987) and subspace (Ephraim and Trees, 1995) methods.”
A Fully Convolutional Neural Network for Speech Enhancement (2016, background): “Decades of works showed feasible solutions which estimated the noise model and used it to recover noise-deducted speech [1, 2, 3, 4, 5].”