Title :
Homomorphic analysis of speech
Author :
Oppenheim, Alan V. ; Schafer, Ronald W.
Author_Institution :
Massachusetts Institute of Technology, Lexington, Mass
fDate :
6/1/1968 12:00:00 AM
Abstract :
Classes of systems which satisfy a generalized principle of superposition have been previously proposed and termed "homomorphic systems," emphasizing their interpretation as homomorphic (i.e., algebraically linear) transformations. One such class appears suited to the separation of signals that have been convolved. In this paper, an approach to deconvolution of speech, based on these ideas, is discussed.
Keywords :
Cepstral analysis; Cepstrum; Convolution; Deconvolution; Laboratories; Matched filters; Maximum likelihood detection; Nonlinear filters; Oral communication; Speech analysis;
Journal_Title :
Audio and Electroacoustics, IEEE Transactions on
DOI :
10.1109/TAU.1968.1161965