An audio conferencing system in which each participant in the conference generates a monaural signal for transmission on an output channel to the other participants, and receives a plurality of input channels from the other participants. The input channels are individually converted to stereophonic signal pairs by respective filter banks having selected predetermined transfer functions. The stereo channels are combined linearly to provide a stereophonic audio output. The system has an echo cancellation system in which an adaptive filter process is applied to the input channels to generate a combined echo cancellation signal for applying to the output channel.