Sinsy
sinsy.sp.nitech.ac.jp
- Free plan
- Linux
- Web
What Sinsy is
Sinsy generates singing vocals from musical scores using DNN- and HMM-based synthesis. Its website accepts MusicXML score files and provides vocal selection plus settings labeled ALP, VIB, and PIT; Japanese appears as a language option. The system models characteristics such as pitch and timbre, and does not take vocal input. The Tokuda, Nankaku, and Hashimoto Laboratory developed Sinsy. Its about page, available in Japanese, describes the research background and version history. That history lists version 5.1 on 2026/05/01, noting fixes to parameter-panel and smartphone/tablet behavior. It also records compressed MusicXML upload support in version 4.4 on 2023/12/25, and the addition of Japanese DNN vocal f01027j_dnn_beta5 in version 4.5 on 2024/12/25. The listed platforms are web and Linux. The Free online synthesis plan costs nothing. The synthesis page says that generating vocals implies agreement to the site's terms of use.
Who it suits
Sinsy suits people who want to generate singing vocals from MusicXML scores rather than provide vocal recordings. Its Japanese language option and vocal controls may be relevant to users working with those settings.
Good to know, and what to watch
- Accepts MusicXML score files.
- Uses DNN- and HMM-based vocal synthesis.
- Provides vocal selection and ALP, VIB, and PIT settings.
- Free online synthesis plan: free.
- Does not accept vocal input.
- Synthesis page requires agreement to site terms.
- About page is available in Japanese.
Plans
Every plan and priceFree online synthesis
Free plan
The verdict
Sinsy turns MusicXML scores into synthesized singing vocals, with selectable vocals and several labeled controls. Its workflow is score-based rather than vocal-input-based, and use of the synthesis page implies agreement to its terms.
What the maker publishes
- About page language
- The opened about page is available in Japanese and describes Sinsy's research background and version history. source
- Compressed files
- The version history says compressed MusicXML upload support was added in version 4.4 on 2023/12/25. source
- Current version history
- The about page lists version 5.1 on 2026/05/01 and says it fixed parameter-panel and smartphone/tablet behavior. source
- Development
- The about page says Sinsy was developed by the Tokuda, Nankaku, and Hashimoto Laboratory. source
- Founded
- 2009 source
- Headquarters
- Nagoya, Aichi, Japan source
- Input format
- The web page accepts a score file in MusicXML format. source
- Languages
- The synthesis settings on the current web page show Japanese as a language option. source
- Purpose
- Sinsy generates singing vocals from musical scores using DNN- and HMM-based synthesis. source
- Recent vocal addition
- The version history says Japanese DNN vocal f01027j_dnn_beta5 was added in version 4.5 on 2024/12/25. source
- Site terms
- The synthesis page says that synthesizing vocals implies agreement to the site's terms of use. source
- Synthesis method
- The about page says the system uses DNN- and HMM-based singing voice synthesis to model characteristics such as pitch and timbre. source
- Vocal control
- The page provides settings labeled ALP, VIB, and PIT alongside vocal selection. source
Checked against the maker's pages on 8 Oct 2026.