| Presentation | 2023-03-14 A Method of Timbre Synthesis Reflecting Impression Using Conditional-VAE Takeru Watanabe, Susumu Kuroyanagi, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In This paper, we aim to propose a method of timbre synthesis based on impressions recalled by humans. We worked on this study with the aim of alleviating the current situation in which designers are forced to rely heavily on their experience and senses when designing tones due to the difficulty of systematically explaining the relationship between tones and human impressions of them. Specifically, we propose a method for generating sound waveforms using a conditional variational autoencoder(CVAE), a type of deep generative model, by further conditioning with impression information. The proposed model consists of two parts: a main waveform generation model that directly generates a waveform for one wavelength, and an impression estimation model that is used as an auxiliary model during training. This two-part structure enables impression conditioning without preparing a large amount of labeled data sets, and is expected to both generate a variety of waveforms and reflect impressions. Finally, verification against changes in the approximate shape of the waveform and listening experiments with subjects are conducted to demonstrate the effectiveness of the proposed method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Timbre synthesis / Conditional variational autoencoder / Impression |
| Paper # | NC2022-106 |
| Date of Issue | 2023-03-06 (NC) |
| Conference Information | |
| Committee | NC / MBE |
|---|---|
| Conference Date | 2023/3/13(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | The Univ. of Electro-Communications |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Brain architecture, General |
| Chair | Hiroshi Yamakawa(Univ of Tokyo) / Junichi Hori(Niigata Univ.) |
| Vice Chair | Hirokazu Tanaka(Tokyo City Univ.) / Hisashi Yoshida(Kinki Univ.) |
| Secretary | Hirokazu Tanaka(NTT) / Hisashi Yoshida(NICT) |
| Assistant | Yoshimasa Tawatsuji(Waseda Univ.) / Tomoki Kurikawa(KMU) / Emi Yuda(Tohoku Univ) / Miki Kaneko(Osaka Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Neurocomputing / Technical Committee on ME and Bio Cybernetics |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A Method of Timbre Synthesis Reflecting Impression Using Conditional-VAE |
| Sub Title (in English) | Conditioning by Impression and Generating Sound Waveforms |
| Keyword(1) | Timbre synthesis |
| Keyword(2) | Conditional variational autoencoder |
| Keyword(3) | Impression |
| 1st Author's Name | Takeru Watanabe |
| 1st Author's Affiliation | Nagoya Institute of Technology(NIT) |
| 2nd Author's Name | Susumu Kuroyanagi |
| 2nd Author's Affiliation | Nagoya Institute of Technology(NIT) |
| Date | 2023-03-14 |
| Paper # | NC2022-106 |
| Volume (vol) | vol.122 |
| Number (no) | NC-425 |
| Page | pp.pp.84-89(NC), |
| #Pages | 6 |
| Date of Issue | 2023-03-06 (NC) |