Free entry, limited seats available
The ANR project ChaNTeR
The aim of the ANR project ChaNTer* was the realization and the musical exploration of high quality singing synthesis systems covering two synthesis modes: a mode "text to singing" in which the system synthesizes sound from an electronic score containing melody and a phonetic description of the text, and a mode "singing instrument" in which the musician uses real-time interfaces to control the vocal synthesizer like an instrument. The project paid particular attention to modeling different vocal styles.
During this afternoon, we will present the main lines of research carried out in the ChaNTeR project, the synthesizers developed - the ISiS software (IRCAM Singing Synthesis), which will be distributed to IRCAM Forum members in March 2018, and two singing instruments: Vokinesis and Cantor Digitalis (distributed under CeCILL license) - and the impact of this research on music creation, through the composition works realized thanks to the results of the project.
*ANR-13-CORD-0011 / Partners: LIMSI (coordinator), Acapela, Dualo.
Program
14h15-14h30 | The ChaNTeR project: introduction, history and state of the art in synthesis of the singing voice
Christophe d'Alessandro, CNRS research director with the group LAM from the Jean-Rond d'Alembert Institute of Sorbonne University
This presentation introduces the ANR project ChaNTeR and gives an overview of the history of the synthesis of the sung voice and its technical and musical issues.
14h30-15h30 | ISiS: singing synthesis and style models
Axel Roebel, Research director IRCAM and Head of the Sound Analysis and Synthesis Team; Luc Ardaillon, former PhD student at IRCAM
The ISiS (Ircam Singing Synthesis) software is the main result of IRCAM's research in the ChaNTeRproject. We will explain the main techniques that are used in ISiS to synthesize singing signals from text and melody, and present the results of our research aiming to control the singing style using models learned from recordings of well known French singers as for example Édith Piaf.
15.45-16.15 | The Opera I.D.: a synthetic voice in the context of musical creation
Arnaud Petit, composer
An artificial singing voice, which is the voice of an artificial character that is protagonist of an opera, is one of the particularities of the opera I.D., composed by Arnaud Petit and written by Alain Fleischer. How this voice / character seeks its place in the context of an opera, is what will be addressed in this presentation. I.D. was created at CDN Théâtre du Nord in October 2017, in collaboration with the CIRMMT, IRCAM, Inria, Omnibus Theater, Le Fresnoy, National Studio of Contemporary Arts and Ensemble Les Siècles.
16h15-16h45 | The composition of vibrato: methods of representation and transformation
Marta Gentilucci, composer
The extension of the singing voice was at the heart of the research residency of Marta Gentilucci in 2017 at Ircam. A central topic were the micro / macro time evolution of vibrato, tremolo and more advanced singing techniques. The presentation will show the compositional ideas behind the research and sound examples of the techniques developed.
16h45-17h15 | Singing voice instruments
Christophe d'Alessandro, CNRS research director with the group LAM from the Jean-Rond d'Alembert Institute of Sorbonne University
The voice synthesis instruments were the second part of ChaNTeR project research. Two singing voice synthesis instruments, Vokinesis and Cantor Digitalis (1st prize in the Margaret Guthman competition for new musical instruments in 2015), will be presented. Examples of concerts will accompany the presentation.