Send to

Choose Destination
J Speech Lang Hear Res. 1997 Jun;40(3):646-65.

Acoustic cues to place of articulation and the McGurk effect: the role of release bursts, aspiration, and formant transitions.

Author information

University of Arizona, Tucson, USA.


The McGurk effect demonstrates that the perceived place of articulation of an auditory consonant (such as/bi/) can be influenced by the simultaneous presentation of a videotape of a talker saying a conflicting consonant such as /gi/. Usually, such a presentation is perceived by observers as "di" or "delta i" (known as fusion responses). The reverse pairing (auditory /gi/ paired with a visual /bi/) results in "bgi." percepts. These are known as combination responses. In the current study, three experiments examined how acoustic information about place of articulation contained within the release bursts, aspiration, and voiced formants and transitions of a consonant contribute to the McGurk effect. In the first experiment, the release bursts and aspiration were deleted from the acoustic signal. This manipulation resulted in a smaller impact on McGurk "fusion" tokens relative to the McGurk "combination" tokens. This asymmetry may be related to the perceptual salience of the release bursts and aspiration for velar compared to the bilabial tokens used in this experiment and their importance for obtaining the combination percept. In Experiment 2, the release bursts and aspiration were increased in amplitude. Results revealed either no effect or a stronger McGurk effect for the manipulated tokens than for the intact tokens. This findings suggests that the McGurk effect for fusion tokens does not occur simply because the release bursts and aspiration are weak. In Experiment 3, low-pass filtering the second and higher formants and transitions was associated with the largest overall impact on the McGurk effect. This suggests that dynamic information contained within these formants is of primary importance in obtaining the McGurk effect. These cues are, however, context-dependent and vary as a function of talker and vowel context.

[Indexed for MEDLINE]

Supplemental Content

Loading ...
Support Center