The grottes d'Arcy-sur-Cure are caves in France with prehistoric cave paintings. Some archeologists believe the paintings were located in parts of the cave that have particularly interesting acoustics. It's unclear why the paintings were placed in these locations; perhaps a sonic or performative element accompanied these early works.
In Grottes d'Arcy-sur-Cure, the acoustics of the caves are modeled and applied to the voices of a choir. The score, written by composer Gabriel Bolaños, used exclusively unpitched extended vocal techniques such as fricative consonants ("ff" "sh" "s" "th") and vocal fry. Conventional singing was omitted in order to avoid any allusion to traditional choral music. The goal was to use only unvoiced utterance as musical material.
The voices were processed and mixed live across 6 speakers throughout the hall. The acoustic models were created from recordings made in the caves. Prof. Jonathan S. Abel, researcher at Stanford's Center for Computer Research in Music and Acoustics (CCRMA), processed the recordings to produce the acoustic models. Additional material support was provided by Marcus Noisternig and the Acoustic and Cognitive Spaces group at the Institute for Research and Coordination in Acoustics/Music (IRCAM) in Paris, France.
Grottes was performed by the UC Davis Early Music Ensemble, with conductor Will Cooper, on May 16, 2015. A stereo mix of the recording is available below. A 5.1 surround recording of the performance is available to download.
Recordings of balloon pops were collected at several locations in the caves. These "impulse response" recordings allow the intensity and density of the acoustic reflections from the cave walls to be analyzed and modeled. A convolution reverb filter is the result, which can be applied to any sound, even live. Jonathan S. Abel has used a similar technique to model the acoustics of the Hagia Sofia, allowing a chorus to perform a Byzantine chant with the acoustics applied, without needing to be in the space itself. For the caves, the musical material originally performed there is unknown.
This project was supported by the Margrit Mondavi Fellowship.