The performance of different synthesis signals in acoustic models of cochlear implants

Abstract

Synthesis (carrier) signals in acoustic models embody assumptions about perception of auditory electric stimulation. This study compared speech intelligibility of consonants and vowels processed through a set of nine acoustic models that used Spectral Peak (SPEAK) and Advanced Combination Encoder (ACE)-like speech processing, using synthesis signals which were representative of signals used previously in acoustic models as well as two new ones. Performance of the synthesis signals was determined in terms of correspondence with cochlear implant (CI) listener results for 12 attributes of phoneme perception (consonant and vowel recognition; F1, F2, and duration information transmission for vowels; voicing, manner, place of articulation, affrication, burst, nasality, and amplitude envelope information transmission for consonants) using four measures of performance. Modulated synthesis signals produced the best correspondence with CI consonant intelligibility, while sinusoids, narrow noise bands, and varying noise bands produced the best correspondence with CI vowel intelligibility. The signals that performed best overall (in terms of correspondence with both vowel and consonant attributes) were modulated and unmodulated noise bands of varying bandwidth that corresponded to a linearly varying excitation width of 0.4 mm at the apical to 8 mm at the basal channels.

The Sound Sensation of Apical Electric Stimulation in Cochlear Implant Recipients with Contralateral Residual Hearing

“…Three different types of sounds were presented in the experiments as described next and illustrated in Figure S2B. These signals were chosen because the findings of previous studies suggested they were likely to be perceived as similar to constant-rate stimulation on one electrode [8].…”

“…Noise-band carriers are the most commonly used because they seem to provide the most accurate simulation for speech intelligibility modeling [7]. However, some researchers have found that existing simulators are not always accurate models of phoneme perception [8], and they may not reproduce exactly the sound perceived by CI users [7].…”

Simulating the dual-peak excitation pattern produced by bipolar stimulation of a cochlear implant: Effects on speech intelligibility

“…Based on a simpler model, Bonham and Litvak (2008) also argued that BP should produce a broader spread than MP when the two peaks are considered. Two previous vocoder studies (Bingabr et al, 2008;Strydom and Hanekom, 2011b) that simulated the excitation spreads of MP and BP partly based their choice of filter bandwidth on saline tank measurements reported by Kral et al (1998). In both cases, the choice of the BP bandwidth was based on measurements corresponding to a single peak of the BP pattern.…”

Effects of envelope bandwidth on importance functions for cochlear implant simulations

“…However, insuring correspondence between vocoder simulations and actual implant performance remains difficult. Strydom and Hanekom (2011) compared the performance of a tone vocoder (resembling the one used here) and eight other vocoder models on consonant and vowel recognition tasks with performance measured for Nucleus CI users with SPEAK or ACE speech processors (Pretorius et al, 2006). Although each vocoder emulated some aspects of CI listening well, none were able to closely approximate all aspects of CI listening.…”