Summary: ROBUST SPEECH SEPARATION USING TIME­FREQUENCY MASKING
Parham Aarabi, Guangji Shi, and Omid Jahromi
The Artificial Perception Laboratory
University of Toronto
10 Kings College Road,Ontario, Canada, M5S 3G4
fparham@ecf, guangji@comm, omidj@controlg.utoronto.ca
ABSTRACT
A multi­microphone time­frequency speech masking tech­
nique is proposed. This technique utilizes both the time­
frequency magnitude and phase information in order to es­
timate the Signal­to­Noise Ratio (SNR) maximizing mask­
ing coefficients for each time­frequency block given that the
direction (or alternatively, the time­delay of arrival) of the
speaker of interest is known. Using this masking algorithm,
speech features (such as formants) from the direction of in­
terest are preserved while features from other directions are
severely degraded. Digit recognition experiments indicate
that the proposed technique can result in a substantial in­
crease in the digit recognition accuracy rate. At 0dB, for
example, the proposed technique results in a digit recog­

Source: Aarabi, Parham - Department of Electrical and Computer Engineering, University of Toronto

Collections: Computer Technologies and Information Sciences