The McGurk effect

Their ultimate conclusion stated that when the talker's face was in view there was a "4dB improvement in the speech reception thresholds of normal listeners" (Watkins, 2003). Walden et al (1975) investigated the multimodal nature of speech perception employing participants whom all had been suffering from mild to moderate hearing impairments for a prerequisite number of years. In quiet surroundings the participants were required to listen to a pre-recorded spoken message. The average percentage of words identified correctly was only 10%. Yet when the message was presented alongside a synchronised face speaking the words up to 90% of the words were correctly identified. The overall percentile range of words correctly identified was vast starting at 20% and continuing to 90%. These studies show that the benefits gained from lip reading are limited depending on the individual's capacity to lip read and the speech to noise ratio. Lip reading abilities of both normal hearing and impaired hearing listeners are more or less of the same overall standard.
             Many other studies have tried to investigate the McGurk effect. Green et al. (1990) used a female mimic alongside a taped male voice. The findings not only supported the previous findings but revealed that the process of perceptual integration is an indiscriminate one. This robustness of the McGurk effect was also tested by other studies. Repp et al. (1983) discovered that varying the mimic's timing and by alerting the participants to the existence of the McGurk effect had little impact on the findings. Whalen et al. (1996) found the McGurk effect continues to occur even when the majority of the visual information is removed through only presenting three video frames of the talker's mouth being shown to the listener.
             Pisoni and Luce (1986) believed the McGurk effect to be a simple bias effect. Fowler and Dekle (1991) attempted to investigate this assertion.