Image data is encoded using a buffer memory, and code amount of image data occupying the buffer memory upon coding is detected and buffer information related to the detected code amount is outputted. The coded image data and the buffer information are recorded on a recording medium. In a case where additional recording or overwriting is started, the buffer information in a position immediately before a start position of the additional recording or overwriting is read out, and coding is started based on value, thereby buffer overflow can be prevented.