Song practice support device

A control method for a song practice support device, comprising:

• an association step of associating model voice data stored beforehand and input learner voice data with each other in a time axis direction;
• a storage step of storing delimiter information representative of delimiter positions defining predetermined sections of the model voice data in the time axis direction;
• a section length ratio calculation step of associating the model voice data with the learner voice data on a section basis in accordance with the delimiter information stored in said storage step and the result of association in said association step and of calculating, on a section basis, a ratio between the section length of a model voice represented by the model voice data and the section length of a learner voice represented by the learner voice data;
• a correspondence point identification step of identifying, on a frame basis with a frame having a predetermined time length, correspondence points between the model voice and learner voice for respective sections in accordance with the ratio calculated in said section length ratio calculation step;
• a pitch shift step of shifting a pitch of the learner voice represented by the input learner voice data so as to be coincident with a corresponding pitch of the model voice data in accordance with a result of association in said association identification in said correspondence point identification step;
• a time stretching step of compressing or extending a predetermined section of the learner voice data in the time axis direction so that a section length thereof is made to coincide with a corresponding section length of the model voice data in accordance with the result of association in said association step;
• and
• an output step of outputting to sound emission means the learner voice data shifted in pitch in said pitch shift step and compressed or extended in the time axis direction in said time stretching step,
• wherein in accordance with the result of identification in said correspondence point identification step, the pitch of the learner voice in respect of a frame in which the pitch of the learner voice corresponding to the pitch of the model voice is detected is shifted in said pitch shift step so as to be coincident with the corresponding pitch of the model voice, and interpolation is performed between the pitches of frames close to a frame in which the pitch of the model voice is not detected to thereby determine the pitch of the model voice for that frame, and
• wherein said steps are implemented at least in part by a computer.

A non-transitory computer-readable storage medium storing a program, that when executed by a computer, causes the computer to execute a control method for controlling a song practice support device, the control method comprising:

• an association step of associating model voice data stored beforehand and input learner voice data with each other in a time axis direction;
• a storage step of storing delimiter information representative of delimiter positions defining predetermined sections of the model voice data in the time axis direction;
• a section length ratio calculation step of associating the model voice data with the learner voice data on a section basis in accordance with the delimiter information stored in said storage step and the result of association in said association step and of calculating, on a section basis, a ratio between the section length of a model voice represented by the model voice data and the section length of a learner voice represented by the learner voice data;
• a correspondence point identification step of identifying, on a frame basis with a frame having a predetermined time length, correspondence points between the model voice and learner voice for respective sections in accordance with the ratio calculated in said section length ratio calculation step;
• a pitch shift step of shifting a pitch of the learner voice represented by the input learner voice data so as to be coincident with a corresponding pitch of the model voice data in accordance with a result of association in said association identification in said correspondence point identification step;
• a time stretching step of compressing or extending a predetermined section of the learner voice data in the time axis direction so that a section length thereof is made to coincide with a corresponding section length of the model voice data in accordance with the result of association in said association step;
• and
• an output step of outputting to sound emission means the learner voice data shifted in pitch in said pitch shift step and compressed or extended in the time axis direction in said time stretching step,
• wherein in accordance with the result of identification in said correspondence point identification step, the pitch of the learner voice in respect of a frame in which the pitch of the learner voice corresponding to the pitch of the model voice is detected is shifted in said pitch shift step so as to be coincident with the corresponding pitch of the model voice, and interpolation is performed between the pitches of frames close to a frame in which the pitch of the model voice is not detected to thereby determine the pitch of the model voice for that frame.