SPEECH ENCODING DEVICE, SPEECH DECODING DEVICE, SPEECH ENCODING METHOD, SPEECH DECODING METHOD, SPEECH ENCODING PROGRAM, AND SPEECH DECODING PROGRAM.

摘要

With respect to a signal represented in a frequency domain, a linear prediction analysis is performed in the frequency direction according to a covariance method or an autocorrelation method to obtain a linear prediction coefficient, filter strength is adjusted to the linear prediction coefficient obtained, and then the time envelope of the signal is transformed by filtering the signal in the frequency direction with the coefficient after adjustment. Thus, in a band extension technique in the frequency domain represented by SBR, pre-echo/post-echo which may occur is reduced without a significant increase in bit rate, whereby the subjective quality of a decoding signal can be improved.