4 METHOD FOR FOUR-PART CHORAL MUSICAL NOTE CREATION USING LANGUAGE MODELING AND SYSTEM THEREOF

摘要

The present invention relates to a method to generate a four-part choral note using language modeling, capable of efficiently performing modeling learning with a small amount of data, and a system thereof. According to the present invention, the method comprises: a step (1) of inputting a melody including composition information of at least one part of four parts; a step (2) of generating a candidate note including a chord in accordance with the composition information of the melody; a step (3) of training language modeling with input data having composition-independent characteristics including components independent to composition; a step (4) of comparing the composition-independent characteristics of a chord predicted through trained language modeling with the characteristics of the generated candidate note to assign a score and select the candidate note with the highest score; and a step (5) of repeating the step (4) to generate the four-part choral note from the selected candidate note. Moreover, the system comprises an input unit, a candidate node generation unit, a candidate note selection unit, and a four-part choral note generation unit.