A speaker system has an acoustic diaphragm and an image projection screen provided on a surface of the diaphragm integrally therewith so as to enable the user to enjoy both realistic image on the screen and powerful sound reproduced through the diaphragm. The rear surface of the diaphragm is closed at a peripheral portion thereof while the central portion of the rear surface of the diaphragm is opened to the exterior, so that the speaker system can provide a high level of reproduced sound pressure in low-pitched region, while the thickness of the speaker is reduced remarkably. The diaphgram can be driven by a plurality of driving units so that the speaker system can operate only with an ordinary stereophonic amplifier. The diaphragm can be composed of a pair of diaphragm elements capable of vibrating independently of each other and the screen may be integrally and commonly attached to the surfaces of these diaphragm elements in such a manner as to allow the diaphragm elements to vibrate independently of each other, whereby the speaker system can operate in stereophonic mode. Any variation in the acoustic characteristic attributable to variation in the condition of installation is compensated for by a change in the ratio of the opened area of the rear surface of the diaphragm to the whole effective vibration area or by a selective provision of a sound absorbing member on the rear side of the cabinet.
