Comparison of Adpative Algorithms for the Application of Echo Cancellation

This work presents an adaptive algorithm-based system identification scheme-based echo canceller. The algorithms take into account FIR filters whose taps are selected using a stochastic gradient-based technique to reduce the signal error obtained from the system. Matlab is used to discuss and simulate the following adaptive filters: traditional LMS, LMF and LMMN, which is a mixture of LMF and LMS. Multiple delayed and attenuated replicas were added to a vocal input in order to approximate the echo. The mean-square error (MSE) was taken into consideration when comparing the algorithms.