FPGA Central

Can a DSP system help me? ... I have been diagnosed with severe tinnitus
(ringing in the ears) and hyperacusis (sensitive to loud noises). After
attending a play last night with earplugs in to prevent further injury and
pain, and not being able to hear anything anyone was saying, I thought there
must be a better way.

I have some experience with microcontroller programming and with simple
audio circuits and understand the concept of digital processing, but am
certainly not up to speed on it. This all leads me to ask and start the
research ...

Can a DSP device be programmed to simultaneously provide : ???
- Compression - To amplify low level a bit and bring down the high-level
- Absolute max noise limiting - To protect from further damage
- white noise generation (2 or 3 seperate, tuneable white-noise
generators would be desired. One way to deal with severe tinnitus is to
"mask" the noise with white noise of matching frequencies, then use sound
augmentation, "a-la-hearing aid" to bring the normal sound up a few db to
make it heard. These are used long-term in an effort to retrain the ear to
ignore the tinnitus noises.
- multi-frequency equalization (6 band or so) - to adjust for limited
range hearing damage.

I cannot find a device out there that contains all of these components in a
single headphone and I need to know if I should pursue building my own. I
have had this problem for many years and it is getting worse, so I don't
mind putting a bunch of time, money and learning into this project to help
myself.

Thank you in advance for any comments, tips, or suggestions you might offer.

Jim
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
=-
Jim Frohoff, jfroh_NO_SPAM_@_NO_SPAM_earthlink.net

Jim Frohoff wrote:
> Can a DSP system help me? ... I have been diagnosed with severe tinnitus
> (ringing in the ears) and hyperacusis (sensitive to loud noises). After
> attending a play last night with earplugs in to prevent further injury and
> pain, and not being able to hear anything anyone was saying, I thought
> there must be a better way.
>
> I have some experience with microcontroller programming and with simple
> audio circuits and understand the concept of digital processing, but am
> certainly not up to speed on it. This all leads me to ask and start the
> research ...
>
> Can a DSP device be programmed to simultaneously provide : ???
> - Compression - To amplify low level a bit and bring down the
> high-level

Easy.

> - Absolute max noise limiting - To protect from further damage

Easy - a non-linear version of the above can do this.

> - white noise generation (2 or 3 seperate, tuneable white-noise
> generators would be desired. One way to deal with severe tinnitus is to
> "mask" the noise with white noise of matching frequencies,

That wouldn't work, since white noise is random. Hmm, or would your brain
just learn to ignore that frequency band? If it does, then why can it not
ignore the tinnitus? My undrstanding of the ear and related systems falls
down here :-)

> then use sound
> augmentation, "a-la-hearing aid" to bring the normal sound up a few db to
> make it heard. These are used long-term in an effort to retrain the ear
> to ignore the tinnitus noises.

Ahh - I see :-)

> - multi-frequency equalization (6 band or so) - to adjust for limited
> range hearing damage.

Yep. This can be easily achieved in the fourier domain.

> I cannot find a device out there that contains all of these components in
> a single headphone and I need to know if I should pursue building my own.
> I have had this problem for many years and it is getting worse, so I don't
> mind putting a bunch of time, money and learning into this project to help

> myself.

I can't see any fundamental problems with this approach - I suspect that the
tweaking to make it just right would be hardest.

Anyway - those in the know will follow this up with more detailed and
positive advice, I suspect.

Ben
--
I'm not just a number. To many, I'm known as a String...

"Ben Pope" <spam@hotmail.com> wrote in message
news:blqc2m$f8d15$1@ID-191149.news.uni-berlin.de...
> Jim Frohoff wrote:
> > - white noise generation (2 or 3 seperate, tuneable white-noise
> > generators would be desired. One way to deal with severe tinnitus is to
> > "mask" the noise with white noise of matching frequencies,
>
> That wouldn't work, since white noise is random. Hmm, or would your brain
> just learn to ignore that frequency band? If it does, then why can it not
> ignore the tinnitus? My undrstanding of the ear and related systems falls
> down here :-)
>

I believe he means that the white noise only needs to be around the
frequency he is trying to mask. I believe it would be white noise band
passed filtered to the size of what's called a critical band (roughly
1/3-octive) with the center of the band matching the frequency of the ring
he's trying to mask. This way other sounds of different frequencies aren't
affected anyway near as much.

So I believe the answer is that you need to filter the white noise source to
a tunable bandwidth. (Filtered 'white noise' I believe is referred to as
'pink noise', but I digress). This should be very doable.

In general what you're asking for probably couldn't be home-made as a normal
sized hearing aid.

Good Luck.

> That wouldn't work, since white noise is random. Hmm, or would your brain
> just learn to ignore that frequency band? If it does, then why can it not
> ignore the tinnitus? My undrstanding of the ear and related systems falls
> down here :-)

Thanks for the reply -

Here's a ref to TRT treatment for tinnitus that utilizes the white noise
masking : http://www.epleyclinic.com/trt.htm

I have located several references to generating white or pink noise with
DSP, so it looks do-able, though this will obviously be a huge amount of
learning ahead for me. My guess is that I just create the white noise and
then setup my two or three bandpass filters to let through what I need to
mask the tinnitus frequencies.

Any suggestions to DSP kits / software that come with some of the code
libraries I will need to start learning how to do the compression, limiting,
eq and white noise generation? I have just spent the afternoon hours
reading everything I can find on the TI dspvillage.com site.

Jim

> I believe he means that the white noise only needs to be around the
> frequency he is trying to mask.

Precisely

> > So I believe the answer is that you need to filter the white noise
source to
> a tunable bandwidth. (Filtered 'white noise' I believe is referred to as
> 'pink noise', but I digress).

Pink noise has even power distribution per octave. So I guess what I need
is just a small bandwidth of "random" noise, whether it's pink or white
probably won't matter. I will be doing some testing shortly, but I believe I
only need 1/3 octave centered on the tinnitus frequencies, as you described
above.

> In general what you're asking for probably couldn't be home-made as a
normal
> sized hearing aid.

No doubt impossible for "normal sized" hearing aid. What I have in mind is
a pocket module that can help make my life a bit more normal again :-)

> Good Luck.

Thanks, I'll need it. Jim

Jim Frohoff wrote:

>>I believe he means that the white noise only needs to be around the
>>frequency he is trying to mask.
>
>
> Precisely
>
>
>>>So I believe the answer is that you need to filter the white noise
>
> source to
>
>>a tunable bandwidth. (Filtered 'white noise' I believe is referred to as
>>'pink noise', but I digress).
>
>
> Pink noise has even power distribution per octave. So I guess what I need
> is just a small bandwidth of "random" noise, whether it's pink or white
> probably won't matter. I will be doing some testing shortly, but I believe I
> only need 1/3 octave centered on the tinnitus frequencies, as you described
> above.
>
>
>>In general what you're asking for probably couldn't be home-made as a
>
> normal
>
>>sized hearing aid.
>
>
> No doubt impossible for "normal sized" hearing aid. What I have in mind is
> a pocket module that can help make my life a bit more normal again :-)
>
>
>>Good Luck.
>
>
> Thanks, I'll need it. Jim
>
>
I think you're on the right track. First make it work, then make it
pretty. If I wanted to do that -- and one of these days I might -- I'd
start with a low-power stereo board and headphones that look like
gun-range ear protectors (and act like them when unpowered), with
microphones mounted right on them. Once you get it tinkered with enough
to be worth using, you can power the board from three sealed lead-acid C
cells at least long enough to enjoy a play or a movie. If you decide
it's worth doing, you could miniaturize it later. You might want to talk
to the people at Songbird. They're always looking for new niches for
their products. If you were one of their guinea pigs, you'd get their
aids free. If it solved only one of your problems, it could still be
worth while. http://www.songbirdhearing.com/

Several people who make boards you can use read and post here. I would
use a 'C33 DSK because I already have one and I don't think it would
compromise what I could achieve. I like the idea of floating point for
one-offs and prototypes, even where fixed point might be better in
production. There are likely to be fewer gotchas for a beginner. If you
want to break down the design into parts, write about here. You'll get
plenty of advice, and maybe even a degree of consensus.

Jerry
--
"I view the progress of science as ... the slow erosion of the
tendency to dichotomize." Barbara Smuts, U. Mich.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Jim Frohoff wrote:

>>That wouldn't work, since white noise is random. Hmm, or would your brain
>>just learn to ignore that frequency band? If it does, then why can it not
>>ignore the tinnitus? My undrstanding of the ear and related systems falls
>>down here :-)
>
>
> Thanks for the reply -
>
> Here's a ref to TRT treatment for tinnitus that utilizes the white noise
> masking : http://www.epleyclinic.com/trt.htm
>
> I have located several references to generating white or pink noise with
> DSP, so it looks do-able, though this will obviously be a huge amount of
> learning ahead for me. My guess is that I just create the white noise and
> then setup my two or three bandpass filters to let through what I need to
> mask the tinnitus frequencies.
>
> Any suggestions to DSP kits / software that come with some of the code
> libraries I will need to start learning how to do the compression, limiting,
> eq and white noise generation? I have just spent the afternoon hours
> reading everything I can find on the TI dspvillage.com site.
>
> Jim
>
>
>
>
>
There are a lot of ways to do this. Analog Devices makes an analog
volume compressor that has nice features. It can be made to output a
maximum level without much distortion. A very fast attack time can
simulate a clipper without generating harmonics the way a clipper does.
The gain for low levels can be limited, so ambient noise isn't amplified
to a distracting level. The noise generators can be analog too, and
summed in, at a point to keep them unaffected by the input level, ot to
rise and fall with it. You can do all of that with the DSP too, but it
may not be as easy for you to tweak: you decide. Start with one channel,
then go DSP stereo once you know what's wanted. It's possible to twist a
pot while you keep listening. With a reasonable program, you have to
recompile or at least load a new table to do the same thing, and A-B
comparison becomes hard.

Jerry
--
"I view the progress of science as ... the slow erosion of the
tendency to dichotomize." Barbara Smuts, U. Mich.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

"Jim Frohoff" <jfroh@earthlink.net> wrote in message news:<dl2gb.3523$Qy2.1014@newsread4.news.pas.earth link.net>...
> Can a DSP system help me? ... I have been diagnosed with severe tinnitus
> (ringing in the ears) and hyperacusis (sensitive to loud noises). After
> attending a play last night with earplugs in to prevent further injury and
> pain, and not being able to hear anything anyone was saying, I thought there
> must be a better way.

You could try to talk to some people I know, who have developed an
active earplug for use in severely noisy environments.

You find them like this:

- Go to www.sintef.com
- Enter "acoustics" in that search box in the upper righ hand side
of the page
- High among the hits (no. two or three this morning) is a link called
"Advanced hearing protection" or something like that.

Go to that page, and you can contact the people mentioned there. I am sure
they would be very helpful in pointing you to resources or references that
you may find useful.

Rune

On Mon, 6 Oct 2003 01:13:08 +0100, "Ben Pope" <spam@hotmail.com>
wrote:

>Jim Frohoff wrote:

>> - white noise generation (2 or 3 seperate, tuneable white-noise
>> generators would be desired. One way to deal with severe tinnitus is to
>> "mask" the noise with white noise of matching frequencies,
>
>That wouldn't work, since white noise is random. Hmm, or would your brain
>just learn to ignore that frequency band? If it does, then why can it not
>ignore the tinnitus? My undrstanding of the ear and related systems falls
>down here :-)

This is one of those areas where the effectiveness of the treatment
depends a lot on the individual, but this sort of thing has been used
with some success on a lot of folks. It doesn't work for everybody,
and some people do self-mask tinnitus without any training (I fall in
that category, fortunately).

Apparently the noise helps to decorrelate the tinnitus in the brain,
which for some people otherwise locks to it like a PLL and amplifies
it (kinda like the "don't think of a purple elephant" trick). Once
the brain gets the hint that the tinnitus can be ignored (after it's
masked with noise for a while), it doesn't have as much trouble doing
it. At least that's my simplified take on what is most likely a very
complex neural processing problem in our wetware. There are some
well-known Tinnitus researchers who claim to know relative
sensitivities, etc., well enough to know exactly how much noise to
apply, etc. It's pretty interesting stuff, but the need is real and
this sort of treatment with level-controlled noise in the frequency
region of interest seems to work for a lot of people.


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Jim Frohoff wrote:
....
> Can a DSP device be programmed to simultaneously provide : ???
> - Compression - To amplify low level a bit and bring down the high-level
> - Absolute max noise limiting - To protect from further damage
> - white noise generation (2 or 3 seperate, tuneable white-noise
> generators would be desired. One way to deal with severe tinnitus is to
> "mask" the noise with white noise of matching frequencies, then use sound
> augmentation, "a-la-hearing aid" to bring the normal sound up a few db to
> make it heard. These are used long-term in an effort to retrain the ear to
> ignore the tinnitus noises.
> - multi-frequency equalization (6 band or so) - to adjust for limited
> range hearing damage.

Hi Jim,

a while back I toured the production floors of a hearing aid
manufacturer here in Switzerland (this tour was organized by the local
section of the AES, the Audio Engineering Society).

They produce devices with microscopic-sized battery-powered
A/D-DSP-D/A-Amp units (about the size of a chick pea) - unbelievable
when compared to what we are used to in pro-audio. The DSPs have a fixed
processing path (something like frequency-crossover, multiband
compressor, 5-band eq and limiter) with programmable coefficients which
should cover your points 1, 2 and 4. It is taylor programmed to your
needs at one of their distributor's lab.

They are also very active in psycho-acoustic research, and each semester
they organize, together with University of Zurich, a course in
psycho-acoustics.

The company's name is Phonak.

Regards,
Andor

I can't add much to this conversation except that I think the Blackfin
running at minimal speed might be the lowest power solution. Even at its
lowest operating speed, it would still be plenty fast for the types of
filtering mentioned. A good front end might be a Wolfson WM8731. This
device is also a low power, high performance device. They are used in
Apple's IPOD. TI makes a very similar part.Jim could simulate everything
with an Analog Devices EZ-Kit and build the final product into a pager box.

--
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

Eric Jacobsen wrote:

> ... some people do self-mask tinnitus without any training (I fall in
> that category, fortunately).
>

Me too. I don't even notice it except when reading material like this
or struggling to hear sounds at very low volume.

A broadband masking generator is easy to build, and filtering with
op-amps or active filter chips is nearly as easy. Injecting that with
those foam padded headphones that don't block ambient sound may not be a
good long-term solution, but it is a good way to learn what's effective.
That's the first thing I'll try if at some point I think I need masking.

Jerry
--
"I view the progress of science as ... the slow erosion of the
tendency to dichotomize." Barbara Smuts, U. Mich.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

In comp.dsp, Jerry Avins <jya@ieee.org> wrote:

>I think you're on the right track. First make it work, then make it
>pretty. If I wanted to do that -- and one of these days I might -- I'd
>start with a low-power stereo board and headphones that look like
>gun-range ear protectors (and act like them when unpowered), with
>microphones mounted right on them.

I'd go with the new Shure in-ear phones, starting at $99. These
claim even better isolation than traditional closed-back headphones,
and are a lot smaller and lighter, and you don't look like the monkey
in that Napster logo. For more info, click on the E2c link at
shure.com.

>Once you get it tinkered with enough
>to be worth using, you can power the board from three sealed lead-acid C
>cells at least long enough to enjoy a play or a movie. If you decide
>it's worth doing, you could miniaturize it later. You might want to talk
>to the people at Songbird. They're always looking for new niches for
>their products. If you were one of their guinea pigs, you'd get their
>aids free. If it solved only one of your problems, it could still be
>worth while. http://www.songbirdhearing.com/
>
>Several people who make boards you can use read and post here. I would
>use a 'C33 DSK because I already have one and I don't think it would
>compromise what I could achieve. I like the idea of floating point for
>one-offs and prototypes, even where fixed point might be better in
>production. There are likely to be fewer gotchas for a beginner. If you
>want to break down the design into parts, write about here. You'll get
>plenty of advice,

Definitely...

>and maybe even a degree of consensus.

I dunno, we'll see about that.

>
>Jerry
>--
>"I view the progress of science as ... the slow erosion of the
> tendency to dichotomize." Barbara Smuts, U. Mich.
>¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
>

-----
http://mindspring.com/~benbradley

Hi Jim and Jerry

The parametric equzlizer demo that comes with the VC33 DSK has several
features that will be worth looking into. Here are some things you
might be able to start with...

- The analog IO is stereo 48Khz at 20bit resolution.

- Board level drain is typically 100-150mA, but this is running full
tilt. Lower the clock rate and it will go down. Also to keep the noise
down, switching regulators were not used. With the bulk of the current
going to the 1.8V DSP core, you can do the math and figure that a custom
board using more efficient switchers could get the current down into the
50-70mA range without much trouble.

- Though not the best electrets (gotta be cheap) the board does come
with 2 electret microphones.

- The equalizer passband frequencies (10 left and 10 right), gains and
Q's are all adjustable.

- Generating white noise is about 4 cycles (see RAND.ASM). You would
want to feed this into a seperate filter and sum it into the final output.

- Floating point is a kind of poor mans logarithmic, and once you have
got your signal in the log domain it is easy to change the log base.

- This demo also includes an unbelievably simple (legal dpt says I must
say this) *patented* floating point ALU trick that shows how to reduce
the sensitivity of truncation noise (Log Differential Compression or
LDC). In general it shows how to get near CD quality sound with 0
(zero) mantissa bits. This could go a long way toward reducing any
additional oddities you might produce by changing the log base.

- Another interesting thing about LDC is that when it does produce
harmonic distortion, it tends to generate the 2nd harmonic, and none of
the higher orders. This too might be usefull as it can make things
sound more natural.

- A soft clipper is also included that can be turned on/off. Basically
if the signal reaches the clipping point the gain is immediately
adjusted down to avoid clipping. The gain then drifts slowly back up.

Hope this helps,
Keith Larson

Note: For simple analog IO protection the op-amp output goes through 470
ohm resistors. Either use an inline amp or a pair of headphones with
fairly hi-z. Most headphones work pretty good.

+------------------------------------------+
|Keith Larson |
|Member Group Technical Staff |
|Texas Instruments Incorporated |
| |
| 281-274-3288 |
| k-larson2@ti.com |
|------------------------------------------+
| TMS320C3x/C4x/VC33 Applications |
| |
| $150 TMS320VC33 DSK's ARE AVAILABLE NOW |
| |
| TMS320VC33 |
| The lowest cost and lowest power |
| floating point DSP on the planet! |
| 500uw/Mflop |
+------------------------------------------+

Jim Frohoff wrote:

> Any suggestions to DSP kits / software that come with some of
> the code libraries I will need to start learning how to do the
> compression, limiting, eq and white noise generation?

Hi Jim,

you may find C/C++ code for noise synthesis, filters, and also a
limiter at http://www.musicdsp.org/archive.php . There's no ready-
made compressor there, but
http://www.google.com/search?q=+site...u+%22%5Bmusic-
dsp%5D%22+compressor will give you lots of reading on that.


Martin

On Mon, 06 Oct 2003 11:32:45 -0400, Jerry Avins <jya@ieee.org> wrote:

>Eric Jacobsen wrote:
>
>> ... some people do self-mask tinnitus without any training (I fall in
>> that category, fortunately).
>>
>
>Me too. I don't even notice it except when reading material like this
>or struggling to hear sounds at very low volume.
>
>A broadband masking generator is easy to build, and filtering with
>op-amps or active filter chips is nearly as easy. Injecting that with
>those foam padded headphones that don't block ambient sound may not be a
>good long-term solution, but it is a good way to learn what's effective.
>That's the first thing I'll try if at some point I think I need masking.

Jerry,

I think for people like us who naturally self-mask the risk of needing
masking at some point is low (at least I hope so). My impression is
that the people who need masking treatment are the folks who have
difficulty coping at the onset and need help learning how to mask
effectively. Since this can be debilitating in extreme cases the need
for good solutions is genuine.

Excuse me, my tea kettle's ready...wait, I don't own a tea kettle...


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org

Ben Bradley wrote:

> In comp.dsp, Jerry Avins <jya@ieee.org> wrote:
>
>
>>I think you're on the right track. First make it work, then make it
>>pretty. If I wanted to do that -- and one of these days I might -- I'd
>>start with a low-power stereo board and headphones that look like
>>gun-range ear protectors (and act like them when unpowered), with
>>microphones mounted right on them.
>
>
> I'd go with the new Shure in-ear phones, starting at $99. These
> claim even better isolation than traditional closed-back headphones,
> and are a lot smaller and lighter, and you don't look like the monkey
> in that Napster logo. For more info, click on the E2c link at
> shure.com.
...

For final use, I agree. (It's hard to mount the microphones on the
in-ear 'phones, though!) It seems to me that, during the course of
experimenting to get the best configuration, ease on getting them
on and off outweighs appearance and maybe even comfort.

Jerry
.... there is always a well known solution to every human problem
— neat, plausible, and wrong. H. L. Menken
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Jerry, Al, Andor, Rune, Eric, Martin, Keith, BenB, "J", BenP ...

THANK YOU for your replies and suggestions. I mapped out my tinnitus sounds
today using some sound analysis software. It appears that the three main
frequencies that are loud as hell seem to be centered around 11.2 kHz, 11.8
kHz and 13.8 kHz. The higher frequency tone is so high when generated on
the computer that I can barely hear it through my ears, but it is screaming
in my head.

Tests with a wide-band white noise generator are, so far, unsuccessful in
masking the noise. Maybe it's not wide-band enough, the speakers may not be
good enough, or perhaps my ears just aren't responsive enough to let the
masking noise in at the high freqs. I am doing more experiments to try and
find out what sounds might mask the noise.

The "shooter's protection" lead is a good one and I have located an
amplified headset that allows max 82 db into my ears while providing 15db or
so of amplification, so at least I'll be able to hears plays and concerts a
bit while figuring out a plan.

For now I am going to take the simple route of creating my custom-bandwidth
white noise as a high-bitrate MP3 using various tools on my pc, then using a
miniature MP3 player to pump it into my ears full-time. The treatment folks
so far want over $800 U.S. for a custom sound generator per ear, which seems
a bit steep.

Thank you all for your help and suggestions. They have given me many new
ideas and directions to pursue, (one of which is to buy a DSP kit to start
learning).

Jim

"Jim Frohoff" <jfroh@earthlink.net> wrote in message
news:_jKgb.2079$dn6.1420@newsread4.news.pas.earthl ink.net...
> Jerry, Al, Andor, Rune, Eric, Martin, Keith, BenB, "J", BenP ...

> Tests with a wide-band white noise generator are, so far, unsuccessful in
> masking the noise. Maybe it's not wide-band enough, the speakers may not
be
> good enough, or perhaps my ears just aren't responsive enough to let the
> masking noise in at the high freqs. I am doing more experiments to try and
> find out what sounds might mask the noise.


At the listed frequencies of around 11-13 kHz, you may have trouble finding
good speakers or headphones to reproduce those frequencies. bottom of the
line speakers or headphones may not generate as much sould as you're
thinking. Also there's some frequency roll off about 1kHz that naturally
attenuates sound. There are ear shape related 'holes' that naturally occur,
but I'm not sure if they exist in that frequency range and I don't have the
documentation handy to check :/ . If your frequency is in one of them, you
might have better luck with headphones than speakers.

Also to begin to achive masking, you need to approach the volume of
whitenoise roughly equal to volume of sound you're trying to mask. To
achieve full masking you will probably need to be 6-10 dB louder. In your
case this may be excessively loud. A 1/3 octive at that frequency with a
large volume is a lot of energy.

An additional complication is that your ear has coarse volume adjustment.
It acts like a broadband gain reduction on the input sound. The sound is
then broken down by frequency with a part of the ear that has roughly a
20-30 dB dynamic range (40 dB for some lucky & young fools). This
broadband adjustment is transparent to you most of the time. You notice it
when a loud sound catches you off guard it sometime hurts, but if the sound
is ramped up, it will be tolerable. Its because the coarse volume
adjustment can keep up. Anyway, if you use a loud enough masking source,
the coarse volume adjustment will adjust up. The problem is that what
happens is the other sounds begin to fall into your own ear's noise floor.
You will develop a loss of clarity or eventually have a broadband masking
occur.

A couple of points the these are just from own theorizing.

One is that you may want to try masking with something broadband, but with a
little more structure than pure white noise. Something like filtered
waterfall or engine noise as long as its tolerable. The reasoning is that a
nearly pure tone might be too easy to discern from pure whitenoise. It may
be easier to mask the tone with some broadband noise that has atleast some
structure in it to distract your brain from ferreting out the pure tone.

The other issue is possible bad news. Its referring to the coarse
adjustment that your ear makes with sound. Since you're brain corrects for
the adjustment, I don't know how this will affect the tones coming from
tenitus. You might be fighting a losing battle by trying to increase the
sound level. The course adjustment might just offset the increased sound so
that it is the same relative to the ringing no matter what level of sound
you use. If the ringing was lower, the masking might work, but if the
ringing is very loud as you describe, increasing the volume will actually
just make the ringing seem even louder and the masking may never be enough
to cancel the noise. Hope this isn't the case.

-John

Sorry this discussion has drifed off-topic from DSP a bit ... please let
me know if this is irritating anyone and I will drop/move this thread.

>> you might have better luck with headphones than speakers.

Agreed. I'll pick up a set of high-end headphones tomorrow to continue my
experiments.

> One is that you may want to try masking with something broadband, but with
a
> little more structure than pure white noise. Something like filtered
> waterfall or engine noise as long as its tolerable.

I'll try that by digitizing some other sounds, then filtering. I bought a
"sound spa" environmental sound machine yesterday as well, but none of the
sample sounds seem to mask well either. "Babbling brook" comes closest.

> You might be fighting a losing battle by trying to increase the
> sound level. The course adjustment might just offset the increased sound
so
> that it is the same relative to the ringing no matter what level of sound

That seems to be the case so far. I had the masking playing loud as heck
through my little headphones, but the tones ride right on top of them. I am
traveling to Portland next week to meet with a bunch of audiology experts on
this thing and will see what they suggest as well. I'll report back here if
I find any kool digital audio gadgets they make for this sort of thing.

Thanks John! Jim

J wrote:
>
> At the listed frequencies of around 11-13 kHz, you may have trouble finding
> good speakers or headphones to reproduce those frequencies. bottom of the
> line speakers or headphones may not generate as much sould as you're
> thinking. Also there's some frequency roll off about 1kHz that naturally
> attenuates sound. There are ear shape related 'holes' that naturally occur,
> but I'm not sure if they exist in that frequency range and I don't have the
> documentation handy to check :/ . If your frequency is in one of them, you
> might have better luck with headphones than speakers.
>

When I was involved in hearing research we used to buy transducers from
this company: <http://www.knowleselectronics.com/>. The transducers are
designed to fit to a small flexible tube which can be mounted in the ear
canal in various different ways. I know that at frequencies above 10 kHz
it can get very tricky trying to accurately control the absolute SPL
that is delivered to the ear canal - small movements of the transducer
tube etc have a big effect. You also have to drive the transducers
pretty hard at the high frequencies to get a reasonable SPL.

You might also want to take a look at the various transducers etc from
Etymotic: <http://www.etymotic.com/>.

If you want to try headphones initially then Sennheiser heaphones tend
to be popular for hearing research and you should be able to find
calibration data on the WWW for various models.
<http://www.sennheiserusa.com/>

Paul