Vocal Imitations and the Identification of Sound Events

Abstract It is commonly observed that a speaker vocally imitates a sound that she or he intends to communicate to an interlocutor. We report on an experiment that examined the assumption that vocal imitations can effectively communicate a referent sound and that they do so by conveying the features necessary for the identification of the referent sound event. Participants were required to sort a set of vocal imitations of everyday sounds. The resulting clusters corresponded in most of the cases to the categories of the referent sound events, indicating that the imitations enabled the listeners to recover what was imitated. Furthermore, a binary decision tree analysis showed that a few characteristic acoustic features predicted the clusters. These features also predicted the classification of the referent sounds but did not generalize to the categorization of other sounds. This showed that, for the speaker, vocally imitating a sound consists of conveying the acoustic features important for recognition, within the constraints of human vocal production. As such vocal imitations prove to be a phenomenon potentially useful to study sound identification. Notes Guillaume Lemaitre is now at Carnegie Mellon University. 1 http://www.cartalk.com/. For instance, in a recent show: –“So, when you start it up, what kind of noises does it make?–It just rattles around for about a minute. Just like it's bouncing off something. He thinks that it could be bouncing off the fan, but it's not there. …–Just like budublu-budublu-budublu?–Yeah! It's definitively bouncing off something, and then it stops.” 2For a compelling example, see http://www.neurosonicsaudiomedical.com/ 3We did not distinguish between /R/ and // 4The participants' confidence in identifying the cause of the sound was measured (see Table 1); the mean value was 6.07 on a scale ranging from 0 (“I do not know at all what has caused the sound”) to 8 (“I perfectly identify the cause of the sound”). 5The 72 imitations are available at http://pds.ircam.fr/imitations.html 6When a cluster only incorporates imitations of the same referent sounds (e.g., 1), this cluster receives the same name as the referent sounds (e.g., 1). When a cluster incorporates imitations of referent sounds belonging to a similar category (e.g., liquids), the cluster receives the initial of this category (e.g., ). 7And one might assume that these are the cases of efficient imitations. 8Here the coefficients of D p are either 0 or 1 so squaring the distances does not change anything, but the centering matrix is also used in multidimensional scaling where the distances are not necessarily binary.