When any of push-button switches on a handheld controller is pressed in a sound input mode, a video game machine generates and temporarily stores frequency data of a tone corresponding to the depressed switch. When a joystick on the controller is tilted to a predetermined direction, the video game machine changes the generated frequency data according to the amount of tilt of the joystick. It is therefore possible to input various sounds in tone using a limited number of switches. The frequency data stored in the video game machine is read later to be converted into audio signals, and outputted from a speaker incorporated in a CRT display. When a melody based on the inputted sound coincides with a predetermined melody set, the video game machine makes various changes in the progress of the game. For example, a hero character can be warped to a position that is different from the present position, or provided with various items.