A game apparatus includes a CPU, and the CPU evaluates a pronunciation of a user with respect to an original sentence (ES). First, envelops as to a volume of a voice of the original sentence (ES) and a volume of a voice of the user are taken, and the average values of the volumes are made uniform. When the volumes are made uniform to each other, a degree of similarity (scoreA) of distributions of local solutions when the volumes are equal to or more than the average values, a degree of similarity (scoreB) of distributions (timing of concaves/convexes of the waveform) of values of the high or low level indicating whether or not the volume is equal to or more than a value multiplying the average value by a predetermined value, and a degree of similarity (scoreC) of dispersion values (dispersion of concaves/convexes of the waveform) of the envelopes are evaluated by utilizing the respective envelopes. On the basis of these degree of similarities (scoreA, scoreB, scoreC), the rhythm of the pronunciation by the user is evaluate.