METHOD FOR CANCELING ECHOED SYNTHETIC STEREO SOUND, PARTICULARLY FOR REMOVING ECHOED SOUND BY NORMALIZING CROSS CORRELATION BETWEEN AN INPUT SIGNAL AND AN ESTIMATE ERROR SIGNAL WITH THE SUM OF ESTIMATE ERROR SIGNAL POWER OBTAINED BY MULTIPLYING AN AUTOCORRELATION MATRIX AND A PROJECTION ORDER IN UPDATING A COEFFICIENT OF AN ADAPTIVE FILTER

摘要

PURPOSE: A method for canceling an echoed synthetic stereo sound is provided to implement a stereo sound by using an HRTF(Head Related Transfer Function) for implementing a VoIP environment and remove sound echoing in a handsfree environment by using an adaptive algorithm. CONSTITUTION: Each voice signal of participants is mapped according to a spatial mapping method. The mapped voice signals are localized by using an HRTF. The mapped voice signals are allowed to pass through an adaptive filter. The localized signals are received through the first and second speakers. Echoed sound from the first and second speakers to a microphone is received through the microphone. An estimate error signal, namely, a difference between a signal inputted through the microphone and the output signal obtained by localizing the mapped voice signals, is measured and a coefficient of the adaptive filter is changed based on the measured estimate error signal. In this case, the coefficient of the adaptive filter is changed by normalizing cross correlation between the mapped voice signals and the estimate error signal with the sum of the estimate error signal power obtained by multiplying an autocorrelation matrix and a projection order.