PURPOSE: To automatically track a speaker, which can easily change the positions and the number of microphones, without limiting the number of conference participants by detecting light emission through a television camera and storing a light emitting position as a speaker position based on the rotational angle of the television camera in the case of the detection.

CONSTITUTION: Seven microphones 4a-4g, which number is equal to the number of conference participants, are respectively arranged on a table 3 corresponding to the positions of speakers, and the respective microphones are connected to a control part 7 in a main body 1 of the system. In order to store the positions of the respective microphones 4a-4g, the light emission at respective light emitting parts 5a-5g is detected by an image processing part 6 through a television camera 2, which is rotated according to the instruction of the control part 7, and in the case of this detected, a camera rotating angle Φn is stored in a storage part 8 as the microphone position, namely, as the speaker position. In the conference, the camera 2 is rotated only at the stored camera rotating angle Φn based on the stored speaker position under remote control or the like and can photograph the speakers.