A separate-type music performance system has a master audio-visual station and a slave audio-visual station remote from the mater audio-visual station and connected through two communication channel independently of each other; MIDI music data codes and click time data codes are transmitted through one of the communication channels to the slave audio-visual station, and audio-visual data codes and a click signal are transmitted through the other communication channel; when the click signal and click time data code arrive the slave audio-visual station, the clock setter 21e sets an internal clock with the click time data code paired with the click signal, and the MIDI music data code are transferred to an automatic player piano in comparison with the time data and the internal clock, whereby the tones are produced synchronously with the visual images.