Greetings All, The last part of the following quote is exactly the concern I have when it comes to auditory research: **Gunther showed..., in the course of his demonstrating that a phantom image is not physically equivalent to a real source, even if is perceptually equivalent (to whatever extent).** Here’s the scenario: I test cochlear implant (CI) wearers or hearing aid (HA) users in a controlled, laboratory space. Five word sentences are presented, one at a time, through one of several loudspeakers. The remaining loudspeaker or loudspeakers present background noise in the form of pink noise or speech babble. I make adjustments to the CI processor or HA settings, and the wearer/listener demonstrates improved speech comprehension ability (based on percentage of correct words or sentences) in the lab. I have no idea what perception is like for these listeners; I can only measure speech comprehension ability/performance. The HA or CI users re-enters the REAL WORLD only to discover that the new processor settings made a whole mess of everything (intelligibility-wise).
I go back to the drawing board with the grandiose idea that I can create *real-world* stimuli using a popular stereo miking technique. For normal-hearing listeners, the new stimuli are PERCEIVED to sound real. Unfortunately, I don’t know how PHYSICALLY real the stimuli are and, consequently, no idea how the acoustical stimuli will interact with a CI wearer’s head, mic, processor, and (ultimately) implanted electrodes. For simple stimuli, such as a single talker, I can make any number of spectral measurements to ensure that the stimulus is realistic. But for complex, multi-directional stimuli, I would like to know that newly-developed stimuli are physically equivalent to their real-world counterparts. Auditory streaming, scene analysis, timbre, etc. all apply to the hearing-impaired, but it is THEIR perception. I’m not actually concerned that their perception is different from mine or from anybody else’s. What matters is when I decide to modify electrical pulse-width (current delivered to electrodes), frequency transposition algorithms, envelope detection (e.g. half-wave versus Hilbert transform), number of electrodes receiving stimulus, etc. is that these changes are in response to PHYSICAL reality that will ultimately enable improved CI performance in the REAL WORLD. Making changes based on normal-hearing listeners’ perception isn’t the key to building better acoustic stimuli for hearing research (at least not for studying HA or CI efficacy). I believed Ambisonics was a way of capturing real-world events and bringing them to a controlled, laboratory environment (kind-of requisite for science, oui?). If I have to use HOA to accomplish this, then so be it. But most importantly, I need some means of showing that this proposed laboratory listening environment and stimuli are physically *real* so that it applies to ALL listeners (hearing-impaired or normal listeners) regardless of their perceptual judgments. If listening ability improves in the *improved* laboratory environment, then I have confidence that changes made to CIs and HAs will make a real difference where it counts: Outside of the laboratory. Thanks for many great insights (and for reading my rants regarding hearing science). Kind regards, Eric C. -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://mail.music.vt.edu/mailman/private/sursound/attachments/20130704/09dbf67d/attachment.html> _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
