DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | 남주한 | - |
dc.contributor.author | Kim, HounSu | - |
dc.contributor.author | 김현수 | - |
dc.date.accessioned | 2024-07-25T19:30:55Z | - |
dc.date.available | 2024-07-25T19:30:55Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1045770&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/320582 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 문화기술대학원, 2023.8,[iii, 25 p. :] | - |
dc.description.abstract | Neural MIDI-to-audio synthesis is a task where given note melody of a specific instrument, realistic audio containing appropriate musical expressions is synthesized. Acoustic guitar possesses various performing techniques, which leads to a rich amount of musical expressions. In this work, we propose an end-to-end neural synthesizer based on diffusion-based generative model that could close the gap between MIDI and realistic guitar sound. We take advantage of the solid conditional nature of MIDI-to-audio synthesis task and propose an effective autoregressive continuation algorithm based on inpainting methods that have emerged in diffusion models. Furthermore, due to the lack of MIDI and audio pair datasets on acoustic guitar, we propose a large dataset where audio is synthesized based on virtual musical instruments and pre-train the model on this dataset in the context of transfer learning. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | 뉴럴 오디오 합▼a어쿠스틱 기타 소리 합성▼a디퓨전 기반 생성 모델 | - |
dc.subject | Neural audio synthesis▼aAcoustic guitar sound synthesis▼aDiffusion-based generative model | - |
dc.title | Realistic acoustic guitar synthesis with diffusion inpainting and transfer learning | - |
dc.title.alternative | 디퓨전 인페인팅과 전이학습을 통한 어쿠스틱 기타 소리 합성 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :문화기술대학원, | - |
dc.contributor.alternativeauthor | Nam, Juhan | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.