Currently, regression coefficients are calculated for each frame in delta. If all the input signal is voiced, this calculation is enough only once.
Log in to post a comment.
