Hi all, Is there a best way for eigenvalue estimation in speech case for DCT-lms? I test my algorithm with power normalization as proposed by Farhang, Ali H Sayed and Haykin, my convergence still under the NLMS one and when I increase the number of taps(500 taps) it become difficult to converge. Is DCT-lms useful for AEC. thank you in advance
DCT-lms
Started by ●March 30, 2009