News

Audiophiles almost universally agree that hearing—or "auditory perception" to neuroscientists—improves with practice. That phenomenon would explain why many of us are able to hear differences between audio components that untrained listeners can't hear.

To our dismay, and sometimes, discredit, few scientific studies of this learning process have been conducted or published. By-the-book engineering types frequently dismiss audiophiles' assertions that they can hear minute differences that are imperceptible to most people, a situation that has fueled one of the most contentious and long-running disputes in the audio world: whether such differences are perceptible and quantifiable in double-blind testing.

A study undertaken by researchers at Northwestern University's Institute for Neuroscience may shed some light on the old controversy. Curious about how people learn to localize sounds, researchers Beverly A. Wright and Matthew B. Fitzgerald devised an experiment that allowed them to measure how auditory perception improved with practice—specifically, detecting small differences in sound levels, and small differences in arrival times of equally loud sounds at each ear. The two auditory aspects are major factors in the so-called "head-related transfer function," or ear-to-ear differences that create localization cues.

Volunteers in the study—32 adults between the ages of 18 and 44, all with normal hearing—were first tested, using headphones, to determine their abilities to distinguish level changes in pairs of tones launched simultaneously in both channels. They were also tested for the ability to detect timing changes in equal-level tones. Following the initial tests, the volunteers were divided into two groups, and spent one hour per day over a 10-day period training to improve on one of the two tests.

Their progress was closely monitored by Wright and Fitzgerald to determine how quickly, if at all, they responded to the practice of these listening skills. Improvements in both abilities rose "noticeably" after only two training sessions. Improvements in the ability to distinguish small differences in sound levels continued beyond the duration of the experiment. Improvements in the ability to detect timing differences was far less pronounced; training to improve this ability yielded only slightly better results than no training at all, the researchers reported. From these results, Wright believes that learning to perceive slight timing differences is a much more difficult process. She and Fitzgerald, who work in the Institute's Department of Communication Sciences and Disorders, hope to expand their research to explore how "acoustic learning patterns occur in real-life circumstances," according to Bruce Bower in the October 13 issue of Science News.

The study correlates nicely with an examination by German researchers of improvements in auditory spatial perception as an occupational adaptation. That study compared spatial localization abilities of career orchestra conductors, pianists of approximately equal experience, and non-musicians of similar age and overall hearing ability. The conductors, whose work requires them to spot musicians who are out of tune or step with their comrades, demonstrated vastly better abilities to perceive peripheral sounds than either the pianists or the non-musicians.

Although small in scope, both studies lend credence to audiophiles' assertions that they can hear minor differences—such as those between amplifiers or CD players, devices with extremely small frequency-response discrepancies—and that the ability to do so improves with practice. The studies also call into question the validity of blind tests that do not account for the skills of the participants or provide a meaningful way for them to learn the sonic characteristics of the equipment or recordings being compared.

The Wright/Fitzgerald study, "Different patterns of human discrimination learning for two interaural cues to sound-source location," was published in the October 9 issue of the Proceedings of the National Academy of Sciences.