# Echo Cancellation over Voice Telephone Network

I'm a bit confused about the way echo canellation over a voice telephone network works like in the following figure

My issue is with the fact that the point of the adaptive filter and LMS algorithim is to drive the error e(t) to zero correct? Then surely, Speaker B would hear nothing as the algorithim converges.

• think about what your $e(t)$ represents. – Marcus Müller Dec 5 '18 at 18:32
• @MarcusMüller It repersents the unechoed, original speech of speaker A that I want speaker B to hear right? – AlfroJang80 Dec 10 '18 at 2:46