The goal of this project is to create a high quality system for synthesizing song that can be used by the general public. The system must not only sing vowels, but must also say the words of a song. This type of system does not currently exist for the French language. The synthesizer imagined will work in two modes. The first is "song from text" mode where the user can enter a text to be sung along with a score (times and pitches), and the machine will transform it into sound. The second is the "virtual singer" mode in which the user controls the song synthesizer in real-time via specific interfaces; just like playing an instrument.
To create these tools, we suggest a combination of advanced voice transformation technologies, based on our expertise in synthesis by selection of units and in synthesis of singing by rules, and innovative interfaces for gestural control. The project focuses on capturing and reproducing a variety of vocal styles (e.g. lyrical/classical, popular/song). Beyond the quality tests ordinarily used for vocal synthesis the usability of the systems will be tested from the outset, paying particular attention to the creative aspects they allow (assessment through mini-concerts and mini composition projects that use the new control gestures, a virtual chorus, and/or a virtual soloist). Project partners will use the prototype of song synthesis that will be developed during the project to offer synthesized singing voice and singing instrument products that are currently lacking, or to improve the functions of currently existing products. The project will offer musicians and performers a new artistic approach to synthesized song, new means of creation that make interactive experiences with a sung voice possible.
Project reference: ANR-13-CORD-011-02.