OBJECTIVE: The purpose of the present study was to gather data on the influence of compression ratio and attack and release times of slow-acting front-end automatic gain control (AGC) systems on speech understanding of cochlear implant users in various listening situations. The data should allow evaluation of the usefulness of front-end AGC in body-worn speech processors. DESIGN: Subjects were 12 experienced postlingually deafened adult users of the Med-El Combi-40 multichannel cochlear implant. Six different front-end configurations, including a linear setting, the standard AGC of the Med-El Combi-40 processor, and four slow-acting dual front-end AGCs (use of two instead of one level detector for improved transient handling), were evaluated in two experiments. In experiment 1, tests were performed at 55, 70 and 85 dB SPL, roughly corresponding to soft, medium and loud speech. Experiment 2 was intended to evaluate the quality of transient handling of the six configurations. In this experiment an intense transient ("chink") at 100 dB SPL was spliced onto the beginning of each sentence (presented at 85 dB SPL). RESULTS: At 55 dB SPL subjects performed significantly more poorly with the linear setting than with the AGC settings, but no differences in performance could be found for the AGC settings. At 70 dB SPL subjects showed the poorest performance with the high compression ratio dual front-end AGCs. No differences in performance of the six front-end configurations were found at 85 dB SPL. In the presence of intense chinks, performance of the standard AGC dropped significantly. CONCLUSIONS: The results indicate that slow-acting front-end AGC can be used effectively in speech processors for cochlear implants to expand the range of input levels that are audible for the cochlear implant user, without any need to adjust a processor control, and that incorporation of an additional fast-acting AGC component can improve performance under conditions where intense transients occur.
OBJECTIVE: The purpose of the present study was to gather data on the influence of compression ratio and attack and release times of slow-acting front-end automatic gain control (AGC) systems on speech understanding of cochlear implant users in various listening situations. The data should allow evaluation of the usefulness of front-end AGC in body-worn speech processors. DESIGN: Subjects were 12 experienced postlingually deafened adult users of the Med-El Combi-40 multichannel cochlear implant. Six different front-end configurations, including a linear setting, the standard AGC of the Med-El Combi-40 processor, and four slow-acting dual front-end AGCs (use of two instead of one level detector for improved transient handling), were evaluated in two experiments. In experiment 1, tests were performed at 55, 70 and 85 dB SPL, roughly corresponding to soft, medium and loud speech. Experiment 2 was intended to evaluate the quality of transient handling of the six configurations. In this experiment an intense transient ("chink") at 100 dB SPL was spliced onto the beginning of each sentence (presented at 85 dB SPL). RESULTS: At 55 dB SPL subjects performed significantly more poorly with the linear setting than with the AGC settings, but no differences in performance could be found for the AGC settings. At 70 dB SPL subjects showed the poorest performance with the high compression ratio dual front-end AGCs. No differences in performance of the six front-end configurations were found at 85 dB SPL. In the presence of intense chinks, performance of the standard AGC dropped significantly. CONCLUSIONS: The results indicate that slow-acting front-end AGC can be used effectively in speech processors for cochlear implants to expand the range of input levels that are audible for the cochlear implant user, without any need to adjust a processor control, and that incorporation of an additional fast-acting AGC component can improve performance under conditions where intense transients occur.
Authors: Ingeborg Hochmair; Peter Nopp; Claude Jolly; Marcus Schmidt; Hansjörg Schösser; Carolyn Garnham; Ilona Anderson Journal: Trends Amplif Date: 2006-12
Authors: Michael F Dorman; Louise Loiselle; Josh Stohl; William A Yost; Anthony Spahr; Chris Brown; Sarah Cook Journal: Ear Hear Date: 2014 Nov-Dec Impact factor: 3.570