To solve the problem without user's time and labor, and cost that, since an photographing person is located at a position closest to a speech input device, his or her sound is recorded in a larger volume than that of a person to be photographed, in a speech input and output device such as a video camera.

Voice recognition is performed for input speech data in relation to a data base. The sound volume is controlled by an analog amplifier or a digital amplifier, according to the result of the voice recognition. Alternatively, filtering processing is performed, according to the result of the voice recognition. Alternatively, after the voice recognition, the sound volume of specified speech such as the speech of the photographing person is reduced by performing directivity control according to the result of correlation calculation.