New CI Approach Improves Sound of Music

For people with hearing problems, a cochlear implant can transform their world. The tiny computer chip is surgically implanted into the skull to simulate the work done by the busy hairs in the inner ear to process sounds-and the implants perform very well rendering spoken language. Melody perception, however, remains a challenge. But a new system that adapts cellphone sound processing appears to bring cochlear implant technology closer to offering the best of both acoustical worlds: speech and music.

The research on this new approach will be presented at the Acoustical Society of America's 161st annual meeting in Seattle, Wash. Compared with the conventional cochlear-implant sound sampling strategy, the new scheme significantly improved melody perception. In a test of nine subjects wearing cochlear implants who were asked to identify 10 melodies, results showed 10-20 percent improvement.

Notes lead researcher Fan-Gang Zeng, Ph.D., research director of the Hearing and Speech Lab at University of California, Irvine: "One potential application of this scheme is to one day integrate cochlear implants with smartphones so that future users can not only get better performance, but also seamless communication. Imagine one device that helps you hear and connects all."

In the current cochlear implant pitch-encoding schemes for rendering melody, the original sound signals are significantly altered. This produces potentially detrimental effects on speech perception, which means improvement in hearing music comes at a cost to hearing speech. To overcome this, the new approach takes advantage of spectral constancy, which refers to unaltered tone-quality perception. It is achieved by preserving the spatial position voiced sounds occupy in a given timeframe, while altering the timeframe of pitch cycles. This minimizes distortion of the sound signals of both speech and music.

The presentation, "Using spectral constancy to encode temporal pitch and improve cochlear implant melody perception," by Zeng et al is in the morning session on Friday, May 27 in Grand Ballroom C.

Abstract:http://asa.aip.org/web2/asa/abstracts/search.may11/asa1190.html