Sinsy (Singing Voice Synthesis System) (しぃんしぃ) is an online

Hidden Markov model A hidden Markov model (HMM) is a statistical Markov model in which the system being modeled is assumed to be a Markov process — call it X — with unobservable ("''hidden''") states. As part of the definition, HMM requires that there be an ob ...

(HMM)-based singing voice synthesis system by the

Nagoya Institute of Technology The , abbreviated to Nitech (or in Japanese to 名工大, ''Meikōdai''), is a public highest-level educational institution of science and technology located in Nagoya, Japan. Nitech was founded in 1905 as ''Nagoya Higher Technical School'', then ...

that was created under the

Modified BSD license BSD licenses are a family of permissive free software licenses, imposing minimal restrictions on the use and distribution of covered software. This is in contrast to copyleft licenses, which have share-alike requirements. The original BSD lice ...

Overview

The online demonstrator is free to use, but will only generate tracks up to 5 minutes. The user uploads data in the

MusicXML MusicXML is an XML-based file format for representing Western musical notation. The format iopen fully documented, and can be freely used under the W3C Community Final Specification Agreement. History MusicXML was invented by Michael Good and ini ...

format, which the Sinsy website reads to output a WAV file of the generated voice. Gender factor, vibrato intensity, and pitch shift can be adjusted prior to output. As of December 25, 2015, the official developers of Sinsy were Keiichi Tokuda (Producer and designer), Keiichiro Oura (Design and Development), Kazuhiro Nakamura (Development and Main Maintainer), and Yoshihiko Nankaku. It was originally only in Japanese and English, but Mandarin was later added; the website only supports English and Japanese despite this currently. In 2016, Sinsy stated using the

deep learning Deep learning (also known as deep structured learning) is part of a broader family of machine learning methods based on artificial neural networks with representation learning. Learning can be supervised, semi-supervised or unsupervised. De ...

processing technology DNN.

Products

*Yoko (謡子), Japanese female vocal. She currently has two vocals for the service, both in Japanese, one being a beta and the other a fully released version. *Xiang-Ling (香鈴), Japanese female vocal. An English vocal was added Christmas, 2015. Mandarin was added also to her language capabilities. *Matsuo-P (松尾P), English masculine vocal *Namine Ritsu S (波音リツＳ), a Japanese male vocal currently in beta. Originally produced for UTAU, it released on December 25, 2013. *Unidentified; an unknown vocal in Japanese *Unidentified; a second unknown vocal in Japanese

References

Speech synthesis software Singing software synthesizers {{music-software-stub