A system and method eliminate the drag associated with CPU drag on experienced by combine audio streams from multiple sources. The embodiments receive audio signals from two separate audio sources and apply a MIDI based timing signature to the audio signals. The audio signals are synchronized according to the MTC signature signal and combined, when synchronized, at a summing module for output as a combined audio stream. Some embodiments include monitoring for and publishing a start/stop command signal to users in the collaboration to help synchronize audio signals from the different users.