Harmonics learning system and method
According to one embodiment, provided is a harmonics learning system comprising: a communication unit for providing, to a user terminal, sheet music of at least one first voice part, and receiving, from the user terminal, sheet music of at least one second voice part that excludes the first voice part; a model generation unit for generating a scoring model by using a plurality of rules divided into vertical elements and horizontal elements; and a control unit for marking harmonic scores with the sheet music of the first voice part and the sheet music of the second voice part by using the scoring model.
1. A harmonics learning system, comprising:
a communication unit that provides a user terminal with sheet music of at least one first voice part, and receives from the user terminal sheet music of at least one second voice part that excludes the first voice part;
a model generation unit that generates a scoring model by using a plurality of rules divided into a vertical element and a horizontal element; and
a control unit that marks a harmonic score with the sheet music of the first voice part and the sheet music of the second voice part by using the scoring model.
2. The system according to claim 1 , wherein level of the scoring model is classified according to a number of included rules.
3. The system according to claim 1 , wherein the vertical element includes at least one of a diapason rule, a position rule, a voice-cross rule, and an overlapping rule.
4. The system according to claim 1 , wherein the horizontal element includes a common rule and a melody rule, the melody rule includes at least one of a stepwise rule, a skip rule, an active-tone rule, and a peak rule.
5. The system according to claim 1 , wherein the model generation unit generates the scoring model through reinforcement learning using the rules.
6. The system according to claim 5 , wherein the model generation unit generates the scoring model by performing repetitive learning so that a total sum of scores given according to the rules is a maximum.
7. A method for learning harmonics, comprising the steps of:
providing, by a communication unit, sheet music of at least one first voice part to a user terminal;
receiving, by the communication unit, sheet music of at least one second voice part that excludes the first voice part from the user terminal;
generating a scoring model by using a plurality of rules divided into a vertical element and a horizontal element by a model generation unit; and
marking a harmonic score with the sheet music of the first voice part and the sheet music of the second voice part by using the scoring model.
8. The method according to claim 7 , wherein in the step of generating the scoring model, the scoring model is generated through reinforcement learning using the rules.
9. The method according to claim 8 , wherein in the step of generating the scoring model, the scoring model is generated by performing repetitive learning so that a total sum of scores given according to the rules is a maximum.