In each of photographed L and R viewpoint images, an object area and an object size are detected. A disparity calculator calculates a maximum disparity of the object area between the L and R viewpoint images. A sound image locating unit determines the shift amount of a reverberation sound ratio from a standard ratio based on the object size and the maximum disparity, to give an appropriate depth to sound data. The sound image locating unit also locates the sound data with an appropriate location angle in accordance with the position of a horizontal center of the object area. The processed sound data and L and R viewpoint image data are compressed, and recorded as an AV file on a memory card.