file.newsgroup.med.59321 Maven / Gradle / Ivy
From: [email protected] (Rich Haller)
Subject: ReSound hearing aid theory as I understand it
The following is based on copies I was given of some articles published in
Hearing Instruments. I would appreciate any comments about this and other
'new' technology for hearing aids.
The ReSound system was developed on the basis of some research at AT&T and
appears to take a different approach from other aids. It appears to me that
a new 'programmable' aid like the Widex just uses a more flexible (and
programmable) version of the classical approach of amplifying some parts of
the spectrum more than others and adding some compression to try and help
out in 'noisy' situations.
The major difference in the ReSound approach is that it divides the
spectrum into low and high frequencies (splitting point is programmable),
apparently based on the fact that lots of vowel information can be found in
the low frequencies, while the important consonant information
(unfortunately for me) is in the high frequencies. The two bands then are
treated with different compression schemes which are programable. They have
also developed a new fitting algorythm that builds on what they call
'abnormal growth of loudness'.
This latter is interesting and fits my own personal experience, though I
think the phrase is missleading. What appears to be the case is that as you
exceed the minimum threshold for a person with hearing loss, the deficit
becomes progresslively less compared to normals and by the time you reach
the 'too loud' point the sensitivity curves appear to converge. This means
that if you just boost all sound levels, you are overloading at the high
end for people with hearing losses. Hence what you want is progressively
less amplification as the signal get closer to the maximum tolerable point.
You want to boost low volume sounds more than high and do so potentially
differently for the low and high frequency parts of the spectrum (specially
for someone like me who is relatively normal up to 1000 cps and then falls
off a cliff).
Aids with simple compressors don't descriminate between energy in the low
and high frequencies and can therefor 'compress' useful high frequency
information because of high volume of low frequency components.
Particularly impressive was the ReSound performance with whispered speech
and in simulated restaurant noise situations.
-Rich Haller University of Oregon, Eugene, OR,
USA