Variable rate signal compression has found many applications where there is no serious limitation on delay and the signal parameters are not very susceptible to errors. Methods used to apply variable rate coding usually rely on the redundancies included in the signal.
More
Variable rate signal compression has found many applications where there is no serious limitation on delay and the signal parameters are not very susceptible to errors. Methods used to apply variable rate coding usually rely on the redundancies included in the signal.
Such methods are different in final bit rate, quality of the synthetic signal and computational requirements. This paper presents a novel method for compression of speech signal in a variable scheme. Based on the known linear prediction method, a simple and efficient model is developed in which segments of the speech signal are classified as voiced or unvoiced using the innovative voiced and unvoiced cycle concept.
Manuscript profile
