PURPOSE: To get a listener to perceive sound image of arbitrary position, by providing 2n-set of speakers and equalizers to n-listener and setting the equalizer output so that the sound pressure of both ears from the speakers and can be equal to that from a specific speaker.

CONSTITUTION: Speakers 8a∼8d and equalizers 10a∼10d are provided for listener 13, 14 and a virtual specific speaker 11 is provided at a position localizing the sound image. Impulse response sample trains from the speakers 8a∼8d to both ears of the listeners 13, 14 are measured and those from the speaker 11 are measured. The tap coefficient of convolver is obtained from both the measuring values and given to coefficient devices 151∼15m of a convolver circuit, coupled with delay elements 181∼18m and an adder 28 to realize equalizers 10a∼10d. Thus, the sound pressure of both the ears of the listeners 13, 14 in inputting a signal to input terminals 9 and 12 is the same, and even if the signal is inputted to the terminal 9, the listeners 13, 14 can perceive the sound image at the location of the speaker 11.