Whether or not underflow is being occurred or underflow is highly likely to occur is judged based on an image encoded data amount in a reception buffer at a time of the judgment or a change in the image encoded data amount with time. When the judgment is affirmative, a composite image data piece corresponding to one frame is generated by extracting an image data piece in the frame memory, decoding part of an image encoded data piece in the reception buffer and replacing part of the extracted the image data piece with the decoded part of the image encoded data piece. Composite image data pieces are repeatedly generated such that an occupancy ratio of replaced part of the image data piece increases each time a piece of the composite image data is newly generated, and the composite image data pieces are outputted in order of generation.