2022 Fiscal Year Final Research Report
Mutual Prediction and Adaptation Model of Communication by Integrating Micro- and Macro-structures and Its Application to Music
| Project/Area Number |
19K12288
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 62040:Entertainment and game informatics-related
|
|---|
| Research Institution | Nihon University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2019-04-01 – 2023-03-31
|
| Keywords | 即興演奏 / 旋律生成 |
|---|
| Outline of Final Research Achievements |
Towards a jam session system that enables a human player and a computer to interact with each other through improvisation, we developed technologies that predict human improvisation and generate computational improvisation as follows:
① We developed a method for generating improvisational melodies using a convolutional neural network (CNN). Because a CNN extracts features using hierarchical convolution layers, we achieved to generate Blues-style improvisational melodies by designing convolution layers based on the hierarchical metric structure. ② We applied this method to the JamSketch system, which allows users to play improvisation using melodic outlines.
|
| Free Research Field |
音楽情報処理
|
| Academic Significance and Societal Importance of the Research Achievements |
音楽の生成は,近年,機械学習を中心としたコンピュータ技術が創造性（クリエイティビティ）を獲得できるかを探求する「計算論的創造性」の研究対象の1つとして注目を浴びている.本研究では,即興演奏の旋律をほぼリアルタイムで生成する技術を実現しており,当該分野に一定の貢献をもたらしたと言える.また,この技術を応用した「JamSketch」は,即興演奏ができない非専門家が簡単に即興演奏もどきを楽しむことができ,これまでにない体験をもたらすと期待できる.
|
