A video game apparatus includes an external ROM. This external ROM has a sound memory area storing sound sequence data. The sound sequence data includes a plurality of block sequences for each of three status of "move", "fight" and "stop" of a player object. A CPU detects a status of the player object according to a mode flag, and read a block sequence to a RAM in compliance with the status detected. It is possible to output a series of game melodies in a manner matching the status of the player object by the repetition of using the plurality of block sequences.