The “ventriloquism”, effect refers to the perception of speech sounds as coming from the same direction as the visually observed speaker (even though they are not actually coming from the same direction). Our experiments provide further proof that a necessary condition for this effect is that there be an association between certain critical visual stimuli of the source and the speech sounds. Synchronized movement of the speaker's mouth, associated in the past with speech sounds, is the most important variable. Delay of even .2 sec. between mouth movements of a human speaker, and speech sounds, leads to a large decrease in the “ventriloquism” effect. Results obtained by systematically varying the horizontal angle between the (concealed) sound source and the “visual” source confirm findings of others that strong “ventriloquism” effects can be obtained for an angle as large as 30°. “Ventriloquism” effects can also be obtained for concealed sources placed behind S, though the effect decreases rapidly as the horizontal angle of source placement is changed from a position directly in back of S.

Localization of direction of sound sources was studied as a function of induced head movements. High- and low-frequency noise and pulse stimuli were used. Subjects had had no special training and were blindfolded to remove visual cues. Sources were located to the right and left of the subject, in front and behind, and above and below the horizontal plane through the ears of the subject. Induced head rotation was found to be especially effective in reducing horizontal localization error. Rotation, pivot, and rotate-pivot movements caused a small but significant reduction in vertical localization error for low-frequency noise stimuli.

Visual stimuli associated with a sound source can affect its localization in space—the basis for the “Ventriloquism” effect. The present experiments show that it is possible to obtain such effects over a much wider angle in the vertical median plane than in the horizontal plane. Such “capture” effects are also shown to be larger, the more similar the visual stimuli are to visual stimuli originally associated with that sound source. Attentional variables appear to be critical.

No abstract

No abstract