FIELD: computer technology.
SUBSTANCE: system relates to the field of computer technology for processing of audio data. Echo is removed from an input signal coming from a terminal microphone including a speaker and a microphone; the signal with removed echo is converted into a frequency domain function; a noise signal is assessed, included in the signal with removed echo in the form of a signal generated by removal of echo from the input signal for generation of a signal with assessed noise; the noise signal included in the signal with removed echo is cancelled based on the signal with assessed noise for generation of a signal with cancelled noise; comfort noise is generated based on the signal with assessed noise; a value of the signal with cancelled noise is compared with a threshold value based on comfort noise for each frequency band; comfort noise is overlaid on the signal with cancelled node; and a signal obtained by overlaying comfort noise on the signal with cancelled noise is converted into a time domain function.
EFFECT: provision of echo cancellation without voice signal break.
5 cl, 8 dwg
