Why do we easily distinguish a sound produced by a breaking glass from the sound produced by a shock on a metallic structure, although the spectral content of the two sounds is very close? Why do we easily accept the ersatz of a horse's hooves made by a sound effects engineer knocking coco-nuts together? These questions, representing every-day examples both illustrate the complexity and the pragmatism behind the rather unknown concept related to the: meaning of sounds. This project aims at establishing relations between the structure of sounds, and their impact on human beings. It is based on an interdisciplinary approach associating acoustics, signal processing, psychoacoustics and cognitive neuroscience. The complementarities of these domains make it possible to address this crucial aspect of sound communication.
The project is divided in four parts based on scientific considerations covering a large variety of fields that strongly interact in order to bring forward answers to questions related to the meaning of sounds.
analysis of non-stationary audio signals: the aim is here to develop time-frequency methods able to characterize sounds taking into account fundamental properties of our auditory system. Hence, we strive to associate mathematical methodologies and auditory models within the same concept.
sound synthesis: this domain, which is fundamental in the project, aims at constructing calibrated and ecological stimuli from physical and perceptual considerations of the sound source. We here intend to develop a sound synthesizer defined by physical characteristics of the sound source (material, geometry) and the sound generating action (rubbing, knocking, scraping...). The association of analysis and synthesis gives access to subtle manipulations of natural sounds.
Perception of sounds: in this part the fields of interest are the intrinsic characteristics of the ear and the activity of the primary cortex. More specifically, two fundamental aspects are here considered, i.e. time-frequency masking and the notion of timbre.
The cognitive aspect linked to the listening of sounds: the notion of meaning of sounds in its context is directly dealt with in this part. The use of brain imaging techniques associated to stimuli that are perfectly calibrated and characterized from a physical point of view should give clues to the access of semiotics of sounds. The relations between sound and vision will also be studied in the case of industrial applications and virtual reality.