INTERSPEECH 2013
14thAnnual Conference of the International Speech Communication Association

Lyon, France
August 25-29, 2013

Show Me What You Listen to! Auditory Classification Images Can Reveal the Processing of Fine Acoustic Cues During Speech Categorization

Léo Varnet (1), Kenneth Knoblauch (2), Fanny Meunier (3), Michel Hoen (1)

(1) CRNL, France
(2) SBRI (INSERM U846), France
(3) L2C2, France

An essential step in understanding the processes underlying the general mechanism of perceptual categorization is to identify which portions of a physical stimulation modulate the responses of our perceptual system. More specifically, in the context of speech comprehension, it is still unclear what information is used to categorize a speech stimulus as one phoneme or another. Here we propose to adapt a Generalized Linear Model with smooth priors, already used in the visual domain for estimation of so-called classification images, to auditory experiments. We show how this promising approach can be applied to the identification of fine functional acoustic cues in speech perception.

Full Paper

Bibliographic reference.  Varnet, Léo / Knoblauch, Kenneth / Meunier, Fanny / Hoen, Michel (2013): "Show me what you listen to! auditory classification images can reveal the processing of fine acoustic cues during speech categorization", In INTERSPEECH-2013, 3167-3171.