Concatenative synthesis allows to rebuild singing or speech sentences from recording of diphones (short speech segments composed of a pairs VC, CV, VV, or CC where V is a vowel and C is a consonant). The following exemples use diphone databases recorded by two singers (male and female) in the context of the project.
The singing parameters: pitch (f0) and rhythm (phoneme duration) were extracted from a sentence and were used to rebuild a synthesis sentence from the diphone database.
/*Synthesis with parameters (f0 and phoneme durations) extracted from an original recording:*/
Chorus Digitalis is the first choir in the world composed of synthetic voices controlled in real time by graphic tablets. It is currently composed of the singing instrument from the ChaNTeR project: Calliphony and Cantor Digitalis. Resulting from research on gestural control of voice synthesis initiated in 2005 by Christophe d'Alessandro, Chorus Digitalis has performed for the first time in March 2011 in Vancouver with its designers, Sylvain Le Beux, Boris Doval, Lionel Feugère and Christophe d'Alessandro. /*The current project is led by Lionel Feugère and Olivier Perrotin, as part of their doctoral thesis.*/
Virtual voice lends itself to all kinds of music games, written or improvised, from the accompanying chorus to complex polyphony, through the solo voice singing in the style of North India, baroque Opera or contemporary music.
Click on the picture to watch Chorus Digitalis performances:
The dualo du-touch is an instrument based on the dualo principle.
Click on the picture to watch dualo performances: