For the past 12 years, Dr. Divenyi and Kara Haupt have been conducting research into what is referred to as the "Cocktail Party Effect." This refers to the ability of humans to attend to one voice in a room full of other talkers, as is the case at a cocktail party.

A common complaint of people wearing hearing aids, as well as from other elderly individuals who do not have hearing loss, is that they cannot hear what the person they are talking to is saying in these situations. Even with normal hearing, or with a hearing aid which is turned up, it is not possible for these individuals to attend to one specific voice in the midst of many.

This problem can be very frustrating to those who experience it, and they can begin to feel "left out" or isolated in group situations. In the long term, this can lead to a decrease in the individual's desire to socialize, and may lead to depression. The Department of Veterans Affairs has designated communication difficulties as a priority concern for research in the area of Speech and Hearing, as improvements in the social life of elderly Veterans contributes greatly to their general well-being.

Initial phases of our research into the cocktail party effect were designed to determine whether or not the deficit underlying the problem faced by elderly individuals was in the peripheral auditory system, or a central processing deficit. After extensive testing of basic auditory abilities in a young subject group and an elderly subject group, including tests of spatial, frequency, and temporal dimensions, statistical analysis has revelaed that peripheral hearing loss typically encountered by the elderly (presbyacusis) does not account for the majority of the deficit, and that rather, the deficit must arise from somewhere in the central auditory faculties.

Further research into the area of speech perception and dynamics has lead to the finding that formant frequency tranistions of the type encountered in vowels are a key part of our ability to perceive and understand speech. In studies that tested subjects abilities to perceive and identify vowel analog stimuli, we have found a high degree of correlation between subjects abilities to perceive speech in noise, and their ability to succesfully discriminate vowel-analog stimuli with formant frequency transitions.

Currently, the motivation behind our continued research into the CPE is the future development of a device which would aid wearers in attending to a particular voice in crowded setting. This next-generation hearing aid would depend heavily on the results of our current research, in that the numbers and parameters required for such a device will be the results of our current studies.

home | research | people | writings | contact | links

Speech and Hearing Research © 2004
page design by Alex Brandmeyer

Speech Gesture Movies: download

alignSpondee Package: download

IEEE Sentences: download

IEEE Masker: download

Environmental Sound Files from Curtis Settino


Environmental Sound Files from Curtis Settino part II



“Best” Tokens of 52 Environmental Sounds


Modified Tokens from Freesounds

Sounds and Scenes for Environmental Sound Context Study
.exe file for Contest Study