5aPP13. Using spectral constancy to encode temporal pitch and improve cochlearimplantmelody perception.

Session: Friday Morning, May 27


Author: Duo Zhang
Location: Dept. of Otolaryngol.Head and Neck Surgery and Biomedical Eng., Univ.of California, 364 Med Surge II, Irvine, CA 926971275, fzeng@uci.edu
Author: FanGang Zeng
Location: Dept. of Otolaryngol.Head and Neck Surgery and Biomedical Eng., Univ.of California, 364 Med Surge II, Irvine, CA 926971275, fzeng@uci.edu
Author: Bishnu S. Atal
Location: Univ. of Washington, Seattle, WA 981952500

Abstract:

Despite good speech performance in quiet, melody perception remains a challengeto cochlearimplant (CI) users. Current CI pitchencoding schemes significantlyalter original signals, producing potentially detrimental effects on speechperception. The present study proposes to take advantage of spectral constancyto encode temporal pitch while minimizing signal distortion. Spectral constancyrefers to unaltered timbre perception when the overall level of a signal isroved. For example, the vowel a sounds the same independent of its overalllevel. The idea is to extract and encode pitch by increasing the gain of allchannels proportionally and synchronously according to the pitch period. Thisscheme alters the temporal envelope of voiced sounds but preserves their spectralenvelope in each time frame. It does not alter temporal envelope of unvoicedsounds, making it a good comprise between encoding pitch and preserving spectraland temporal envelopes. The new scheme has been implemented in a realtimespeech processor and evaluated in nine nucleus CI users who identified tenmelodies with minimal training and in a test protocol that randomized thestimulus presentation order. Compared with the standard continuous interleavedsampling strategy, the new scheme significantly improved CI melody perceptionby 10–20%.