|
| A while back Ed made a post (hoh concerns board) about how an aid tuned
for speech may not sound very natural. Well I think I figured out a way
to do this. Basically I took the normal frequency and gain level of all
the sounds in the English language and got my audiologist to program my
aid (acuris) so that these sounds would not be compressed but just
amplified so that they fit in my range of hearing. Compression is used
just to cut off the loud sounds and to process the sounds out of the
range of speech. Well to sum it up I think I can hear much much better
than before.
My solution came about when I came across this
http://www.ihc.com/xp/ihc/documents...earingtest.pdf. Page 3 has a
chart that actually charts out the various sounds used in the English
language. I don't have the exact measurements of the frequency and
gain levels of the sounds so I went with some rough estimates. If
somebody could get some hard numbers my fitting technique can be
improved.
So I've determined that following are the range of sounds I need to
program around
8khz 20 db - 30 db for the "f", "s", "th" sounds
4khz 20 db - 30 db for the "f", "s", "th" sounds
2khz 20 db - 45 db for the "p", "h", "g", "k",
"ch", "sh" sounds
1khz 20 db - 50 db for the "p", "h", "g", "k",
"ch", "sh" and for lower sounds like "r"
500 hz 20 db - 60 db for the rest of the alphabet
250 hz 20 db - 60 db for the rest of the alphabet
As an example of how to fit let's take the 8khz level. I can only
hear between 80 db - 90 db at this frequency. I am assuming that 90db
is the highest level at which I can safely hear. I am lucky since
normal voice occur within a 10 db span so I can fit this sound in my
hearing range without the need for compression. So I would set the gain
on the aid to 80 db - 20 db = 60 db. Then I would set the input
compression knee point to 30 db (had to set to 40db since acuris
doesn't get lower than this) and set the compression ratio to a very
high level like something above 2:1. I set the compression level high
since I don't have much useable hearing at this frequency. The
compression ratio would be lowered for lower frequencies where I have a
greater range of hearing.
At 2khz I can hear between 65 db - 90 db so I have 25 db of useable
hearing range. This is great since I only need 25 db of range for
speech. So I set the gain to 65db - 20db = 45 db. I then can set the
compression kneepoint on the input sound to 45 db (50db on the acuris
since it doesn't go to 45 db) and set the compression to something
like 2:1.
For the lowest frequency (250hz) where my hearing is normal, I have
absolutely no compression and very minimal gain (10db). I don't think
my aid (acuris) will apply the compression on the lower gain levels of
the sound (I heard some aids can do this). If it could I would set the
compression ratio very high and the knee point to 20db to stamp out any
low level sounds that are probably background noise.
For the long time aid users your aid will sound "soft". This will
be because you are used to having low frequency sounds being
overamplified. Please give the above a chance. You will be shocked as
to how well you can hear speech.
Please note that my programming style is very critical of compression.
IMO it's great for hearing anything other than speech. It's the
reason why you can't understand the words of your favorite songs
(Louie Louie is great example on how compression can mangle speech). So
I rather avoid it at all cost when it comes to speech.
Also note that I am using my hearing limitations to eliminate sounds
below speech level. Like I said I only care about improving speech
interpretation. These setting won't be great for listening to music
but it will be great for conversation in noise.
Any suggestions or critiques are welcomed.
George
|
|