The effects of sound coder carrier rate and modulation bandwidth on voice pitch perception in cochlear implant users

Abstract

We employed the dual filter-bank “STEP” coder to separately control the spectral and temporal modulation resolution of analysis channels. Previously we compared vowel pitch ranking and gender classification with eight subjects using enhanced modulation at F0—including across-channel synchronised modulation to the ACE coder. There was no significant improvement using modulation enhanced coding versus ACE across subjects. In a follow-up experiment we looked at the effect of stimulation rate on voice pitch perception. Since there are large inter-subject differences in overall temporal pitch acuity we hypothesised that some subjects’ performance may be more greatly influenced by carrier rate than others, or that some subjects may find sound quality satisfactory with lower carrier rates than those in their clinical processors. We used a version of STEP with a very short temporal envelope analysis window of 2 ms which allows a very low latency real-time processing implementation and large maximum modulation bandwidth. Subjects were tested using carrier rates of 1000, 500 and 250 pps/ch with modulation bandwidths controlled via low-pass filtering. Pilot data indicated that the new low-latency coder provides very good sound quality compared to ACE using 1000 pps/ch or 500 pps/ch. Also the modulation bandwidth could be tuned at different carrier rates to optimize voice pitch perception based on temporal cues. This opens the potential for lower stimulation rates to be used in CI coding while maintaining optimal temporal resolution.We employed the dual filter-bank “STEP” coder to separately control the spectral and temporal modulation resolution of analysis channels. Previously we compared vowel pitch ranking and gender classification with eight subjects using enhanced modulation at F0—including across-channel synchronised modulation to the ACE coder. There was no significant improvement using modulation enhanced coding versus ACE across subjects. In a follow-up experiment we looked at the effect of stimulation rate on voice pitch perception. Since there are large inter-subject differences in overall temporal pitch acuity we hypothesised that some subjects’ performance may be more greatly influenced by carrier rate than others, or that some subjects may find sound quality satisfactory with lower carrier rates than those in their clinical processors. We used a version of STEP with a very short temporal envelope analysis window of 2 ms which allows a very low latency real-time processing implementation and large maximum modulation ban...