Summary of Research Program


For Clinicians and Scientists

The current focus of this research program is to develop better methodology for the measurement of loudness and to apply the methodology to measurement of the loudness of broadband sounds. Abnormal and uncomfortable loudness is a major cause of rejection of hearing aids, despite the fact that many hearing-aid fitting algorithms are specifically designed to restore normal loudness. Much of our understanding of loudness was gained in basic research using narrowband sounds. To study the loudness of broadband sounds, we need tools that will provide information on the contributions of specific frequency regions to the overall loudness of the sound. We are developing those tools using methods widely used in studies of masking, where the impact of specific stimulus properties on perception is determined by computing the correlation between those stimulus properties and subjects’ responses across many stimulus presentations. By varying the level of noise in specific frequency bands from presentation to presentation and having subjects make loudness judgments, it is possible to determine the relative contributions of the stimulus bands to the total loudness of broadband sounds. We are comparing these measures of perceptual weight to data obtained in loudness matching experiments and to perceptual weights predicted by the Moore and Glasberg (2004) loudness model. Preliminary work with the model suggests that the perceptual-weights measure is highly correlated with specific loudness, the contribution of each frequency to total loudness as predicted by the model. Observed perceptual weights data, however, are not highly correlated with the perceptual weights predicted by the model.


Research in psychoacoustics is concerned with the relation between the physical properties of sound, such as frequency and intensity, and the psychological or perceptual properties, such as pitch and loudness. Physical properties of sound can be readily assessed, but measuring the sensory or perceptual experience evoked by a sound is more difficult. Audiologists are required to make such measurements every day, establishing when sounds can be heard, for example, or how loud they are. Our research program is focused on the problem of using the physical properties of sounds to predict the loudness of sounds in listeners with normal hearing and sensorineural hearing loss. We know a lot about loudness, but most of the work has been done with pure tones rather than broadband sounds like speech. When broadband sounds are too loud, one solution is to reduce the level of the entire sound. For listeners with hearing loss, this may make some parts of the sound inaudible. It would be better to know which frequency regions are making the biggest contribution to loudness and to reduce the levels of those specific frequency regions. We are developing the necessary measurement techniques. They may require more test time than is typically available when adjusting hearing aids, but could be quite useful in hearing aid research.