In Voice-over-IP or like systems, audio mixers are used for implanting conferencing applications for IP networks. In such systems, a problem exists as how to pass through user signal events, such a DTMF digits, in-band as well as out-of-band telephony events (RFC 2833/4733) because they are not handled well in such systems. In accordance with the invention, a signalling event in incoming audio packets or user signalling event received out-of-band is detected at one of the inputs. A meta-data representation of the signal event is attached to the audio and passed through the mixer for regeneration, as in-band or out-of-band telephony events, on the other side of the mixer.