To obtain a more suitable grading result, when there are a plurality of singers, as regards singing grading of a karaoke machine.
Pitch differences are calculated (S22, S232) from pitch data of a karaoke singing extracted (218) from a voice signal inputted by a plurality of voice signal input means, and pitch data (S220) of a melody part and pitch data (S230) of one or more harmony parts read out of musical piece data, voice signals, having the least pitch differences, are determined as being the voice signals of the melody part and one or more harmony parts (S226, S236), and pitch differences, corresponding to the melody part and one or more harmony parts, are calculated from the determined voice signals, to grade the musical piece on the basis of the pitch differences.
TAKEMOTO YUKIO
Kenji Takenaka