hearing aids revisited

[MrDEB]

I am curious as to how hard it would be to build a hearing aid such as this
https://www.electro-tech-online.com/custompdfs/2009/11/j2001amicrophone.pdf
200-6khz divided into 12 filters.
Goldmine has the mics but the telicoils unless a "speaker" small enough to fit into ears but operated with out wires to amp.
using several high end op amps for filters and maybe an amp wouldn't be needed seeing how the op amps would boost the volume??
thinking very seriously about ordering some mics from Goldmine.
I wonder if a PIC could be programmed to do all the filters in one chip??

 

[birdman0_o]

It would need to be a high end PIC to be able to do the FFT's

 

[crutschow]

That appears to be a rather complex design that has likely required many hundreds of hours of design and testing to optimize its operation, including the electrical design, signal processing, and the physical layout. It apparently has some complex processing to generate the microphone beam patterns and noise reduction. They designed two custom ASIC chips to minimize power and space.

Unless you are willing to perform a similar effort in its design I would expect your results to be less than satisfactory.

 

[Sceadwian]

I'm sure you could do it with DSP's but like crut said they did theirs with ascics, so you're gonna have power issues.

 

[MrDEB]

I was thinking along the lines of a bandpass filter

like that used for a hi fi stereo.
of coarse use smt. As far as space, the collar looks like it has plenty of room.
Right now my hearing aids work. well one from each pair work (I have two pair).
At $5k a pair I could care less about having ear buds (Ipod) if I could save some money and maybe add to reliability.
Hearing aids have a life expectancy of 3.5 years. yea ouch!!
as far as power, couple of AAA batteries perhaps?
The big reason hearing aids quit working is heat, moisture (sweat) and wax.
with all the electronics on ones chest you may eliminate most.
Reason hearing aids are not moisture proof (as I was told) is the battery needs air to work (reason for little peel tabs on new batteries)
I would like to find a unit that solves most of the headaches involved with available hearing aids. Looking for alternative.
A PIC would be nice, lower the parts count but not a necessary component.
Now designing a good filter maybe another story?
might be educational at best.

 

[crutschow]

If you know your hearing loss response then you can design a filter to provide the inverse of that which, in theory, would give you a normal hearing response. Certainly a multi-channel stereo equalizer, such as used in Hi-Fi audio, should work for that. Getting it small and low power is the problem.

PIC has a DSP µP which may work for this.

Or perhaps something like this 5 band graphic equalizer circuit using a single IC/chip, http://www.datasheetcatalog.org/datasheet/rohm/ba3812l.pdf which uses an analog approach would work. It only uses 5mA and can be ganged to provide more bands. For example, 3 chips can provide a stereo, 7-band equalizer.

Using two directional microphones, one on each end of the chest module, would give some directionality to the hearing. Angling the mikes toward each side should also help.

Edit: Here's an interesting article on directional microphones https://www.electro-tech-online.com/custompdfs/2009/12/DualMic.pdf.

 

[Sceadwian]

Cruts, hearing loss is more complicated than simple response graphs, actually as far as voice distinction goes there's a lot of information in the phase difference from left and right ears, and the primary 'phase encoder' for the human ear is the outer ear structure itself (it's funny lookin for a reason)

This is why the that hearing 'band' that was refrenced uses a microphone array and complex DSP electronics, to attempt to recreate the phase differences needed for the brain's voice recognition section to home in on the components it needs to hear the human voice. It's not about simple directionality because the structure of the human ear has some very complex phase/frequency changes on the audio that comes into it, and directionality isn't the only reason. The outer ear actually forms a tuned waveguide in the forward direction for some frequencies and harmonics, has a lot to do with how we pick up voices; Why do you think the first response any person makes when they hear someone's voice is to look directly at it. Basic time of arival is fine for direction sensing, but the harmonics in the human voice can be seperatly processed, which is why we can hear people over ambient noise in some situations even if their voices isn't.

Otherwise it would be that simple. The problem with hearing aids is in order to work 'properly' they would have to record the sound phase perfect from WITHIN the ear canal. The best ones do. The best ones also cost more than a decent used car. Some advanced 3D systems have used recordings from within a generic human ear mockup with good results, however frequency responce from typical microphones and speakers able to be placed within the ear canal lack.

 

[MrDEB]

I guess I should forget this project idea.

 

[Sceadwian]

Well don't forget about it, it's not a bad idea, it's just quiet a bit to bite off.

 

[MrDEB]

got to get this one

here is a speech filter and a speech communication amp.
Thinking about using a couple of ear buds for output to ears, several SMD mics for input.
Run on several AAA batteries.
BUT not to familiar with analog amps etc. do you think this schematic is feasible?
need to add a volume control and perhaps some SMD pots to adjust the filters?

 

[BrownOut]

For the record, I use very high end hearing aids with advanced speech processors. They really come down to hoity-toity amplifies with a response that is, as suggest, the inverse of my natural hearing loss. They have a directionality feature which uses two mics per unit, and when activated, the signal from the rear mic is inverted for cancel the sound form behind. It works well, but is not a replacement for good natural hearing. The difference between my pricey hearing aids and generic "amplifies" really comes down to the filters that emphasize the frequencies missing from my unaided hearing, aside from some noise reduction. I honestly believe they are little more than over priced amplifers/DSP processors. I imagine that a properly tuned amplifier would give satisfactory hearing augmentation. Me overpriced hearing aids don't really don’t much more than that.

I’ve actually considered replacing them with lower end units, because some of the “advanced” features turn out to by just annoyances.

 

[MrDEB]

I myself have mid-high end hearing aids that have 4 seperate programs that I can change with the push of a button.
These are BTH(behind the ear) and are destroyed by sweat. Come summer=aids take a dump.
At $5000 per pair and average life of a hearing aid is 3.2 years.
I want to build a simple hearing device that runs on AAA batteries and use "ear buds" for output.
Looking at using the SMD mics as well as the schematic I posted. Come the county Fair (I cook breakfast every morning) I get real sweaty over a hot griddle.
Plan is to use this DIY device while cooking or other high sweat task(in my shop)
Use two mics in front on a necklace arrangement and two mics in back. One stereo bandpass filter/amp per channel.
Looking at under $25 complete.
Yes I have tried them cheap SONIC EAR as seen on TV. All they do is amplify ALL the sound. No filters at least none that I could tell.
Maybe by using a speech band pass filter (first two op amps) it might cut down on the mass of sound??

 

[crutschow]

If you know that your hearing loss is at a particular range of frequencies, then you could use a filter to amplify those frequencies (which is what the fancy digital aids do). Perhaps that, along with the speech filter, would be more effective.

 

[MrDEB]

thats why I was planning on using pots in place of set resistors so I could adjust as needed.
Not real up on amps etc but I Y O do you think the amp schematic I posted would be sufficient volume wise?
Maybe we both could build the same unit. LOT cheaper for everyday use than using the high end hearing aids.
Been mulling over a circuit board design as well as parts availability.

 

[audioguru]

Why use a "speech filter" that cuts the very important consonants (s, t, p, th, f, etc) frequencies of speech?
Then mat and nap sound the same.
Then faucet and soffet sound the same.
Then choose and shoes sound the same.
Then faint and paint sound the same.
The difference between the words sailing and failing are in the frequencies from 4kHz to 14kHz.

Reverberation (echoes) and accents compounds the problem of trying to understand speech that has the high frequencies missing.

Without the important high frequencies, your brain gets tired trying to figure out what is being said.

My two cents.

 

[MrDEB]

I will accept your two cents.
Yea I know its in the mail
Good point on the high frequencies.
Perhaps extend the high end of the speech filter or add another filter that adds the higher frequencies then combine using a mixer circuit?

 

[audioguru]

You need an equalizer that produces the same response as your hearing aids. Probably high frequency boost.
You must limit the amount of boost so that too much loudness does not destroy what remains of your hearing or causes feedback.
The amplifier will probably have a limited amount of "headroom" so that the battery can have a low voltage then either you must frequently fiddle with the volume control (to avoid clipping distortion) or have an automatic volume control circuit.

 

[MrDEB]

I had the thoughts of an automatic volume control. See if you can follow, just thinking out loud
a resistor ladder that is connected to a PIC. Then have another pin on the PIC as an ADC input. Then as the volume increases the analog signal increases thus enabling a different resistor in the resistor ladder thus lowering or raising the volume.
REMEMBER I am just thinking out loud.

 

[audioguru]

For many years an automatic volume control has used a Jfet as a variable attenuator resistance. It is linear, not a jerking step-by-step digital attenuator.

 

[MrDEB]

a jfet as a variable resistance?
I am gona do a search on that tomorrow morn.