next up previous contents
Next: Temporal coherence Up: The need for Previous: The need for

Spatial coherence

It has been established that either modality influences spatial localization of the source through the other [20]): Subjects who are instructed to point at a visual source of information deviate slightly from it if a competing acoustic source is heard from another spatial position, and conversely, subjects deviate more from the original acoustic source if a competing optical source interferes from another location. In speech, such a capture of the source is well known and widely used by ventriloquists, as the audience is much more attracted by the dummy whose facial gestures are more coherent with what they hear than those of its animator [354]! Even four-to-five month old infants, presented simultaneously with two screens displaying video films of the same human face, are preferentially attracted by a face pronouncing the sounds heard rather than a face pronouncing something else [165]. This demonstrates a very early capacity of humans to identify coherence in the facial gestures and their corresponding acoustic production. This capacity is frequently used by listeners in order to improve the intelligibility of a single person in a conversation group, when the well-known ``cocktail party effect'' occurs.

Esprit Project 8579/MIAMI (Schomaker et al., '95)
Thu May 18 16:00:17 MET DST 1995