| 摘要 |
PROBLEM TO BE SOLVED: To accurately score and display the similarity of the model singing of a singer or the like and the singing of an amateur singer or the like. SOLUTION: A similarity score device 8 extracts the feature point of the voice of model singing signals Sb and singing signals Sc from a KARAOKE (orchestration without lyrics) reproducing device 1 and performs a specified speaker word voice recognition processing. Voice quality and the length of a generated word obtained by the processing are compared and score video signals Sd for indicating the similarity of the model singing (model singing signals Sb) of the singer or the like and the singing (singing signals Sc) of the amateur singer are outputted from the distance value of a separation state in the comparison. Audio signals in which accompaniment signals Sa, the model singing signals Sb of the singer or the like and the singing signals Sc reproduced by the KARAOKE reproducing device 1 are synthesized are outputted from a speaker 9. Video signals Sv from the KARAOKE reproducing device 1 and the scored result of the score video signals Sd from the similarity score device 8 are screen-displayed at a monitoring device 10. |
