Re: AUDITORY Digest - 17 Jul 2009 to 18 Jul 2009 - Special issue(#2009-163)

[Matt Flax <flatmax@xxxxxxxx>]

Dear Margret and list,

I am a very strong believer in the assessment of individual causes of 
hearing loss ...

Recently Chuck Berlin came to visit us at work ... he pointed out very 
clearly that simply probing an audiogram doesn't really isolate what it 
is which is physiologically dis-functioning ....

I am a firm believer in this approach to diagnosis and cure ... Jont's 
approach to understanding hearing impairment (as per the information on 
your html link below) is most defiantly an appropriate direction to take 
...

Currently we use a prescription to define the nature of non-linear gain 
amplification on an individual basis ... the gain prescription is 
becoming more and more advanced, to the point where the soon to 
be released NAL-NL2 prescription formulae is expected to tune the 
nonlinearity closer to what is clinically expected then ever before.

On a different tangent ....

In my opinion we are on the cusp of the next phase ...

No one has managed to invent it yet, and it is possible that no one will 
... if you talk with those who have been around the industry for quite 
some time, some of them have tried and have emerged on the other side as 
pessimists ...

People are skeptical, more and more of the innovators see the answer to 
the cocktail party noise as being the introduction of more 
directionality ... think about it, more directionality, less peripheral 
stimulation ... only the signal of interest makes it to the 
nonlinearity ... this gets corrected ... etc etc etc 
Well that paradigm works well (when the environment is not 
largely reverberant)

>From a cold engineering perspective, I would say that the majority of 
hearing correction comes from the core algorithm (currently the 
nonlinearity) and everything else tacked onto that is only capable of 
achieving a small improvement ...
To get close to perfect hearing requires more and more complexity, 
possibly with each extra stage adding new artefacts which require their 
own correction ... the mind boggles at how much power a hearing aid DSP 
will eventually require !

So whilst I develop hearing aids which use traditional approaches of 
nonlinear amplifiers and directional microphones/processing, I am most 
certainly setting my sights on the future ... the new approach ...

The new approach must start with a clean slate and re-develop the core 
algorithm ... in my opinion the core algorithm will be the largest 
percentage of correction ... my belief in this has been tempered by the 
pessimists ... they are not willing to support such research and 
development in any way ... it is just too risky in their opinion ...
The only people who are willing to support are those in the hearing aid 
manufacturing business ... even they will only offer in-kind support ...
So we may be waiting a while yet ...

Your suggestion of efferent connections has been a favourite topic of 
mine for getting close to a decade ... it questions the separation of 
place and temporal analysis ...
A great limitation of current hearing aids is that the prescription 
(place nonlinearity) is tempered by the temporal compressors ... just to 
exemplify my concern !

Matt

On Sat, Jul 18, 2009 at 04:43:44PM -0700, Margaret Mortz wrote:
> Jont,
> 
> You have just reminded us of the terrific pioneering research that was
> done at Bell Labs by many, yourself included.  The unfortunate closing
> of the Labs and the spreading out of Bell people may have energized
> many other institutions.
> 
> I went to your website to look at your recent  research on hearing aids.
> http://www.ece.illinois.edu/news/headlines/hl-allen-hearing-aid.html
>  As a long-term hearing aid wearer (childhood measles/mumps; later
> Meniere's), I agree that hearing aids don't work well, especially in
> noisy situations. The latest models have many helpful features, but
> there's still a performance gap.
> 
> If your testing can find the critical cochlear and post-cochlear
> components of an individual's hearing loss, what type of signal
> processing would be needed to compensate?  Phonak and other aids have
> some noise reduction algorithms, adaptive directional mikes, multiband
> frequency channels, and adaptive compression in multiple listening
> programs.  Still something seems missing.  What else is needed?
> 
> I'm also curious on how the olivocochlear connections to the cochlea
> affect the ability of a healthy ear to understand speech in noise,
> specifically "cocktail" party noise..  Are there any connections  fast
> enough to respond to the pitch/prosody of speech that could allow
> real-time cochlear tuning?
> 
> Margaret
> 
> 
> les of the outer hair cells to improve hearing aids.  In a healthy
> ear, how does the feedback back from the brain adapt the coch to allow
> better hearing in a cocktail situation?  Is the
> 
> 
> Assuming you can find electronics that can
> 
> 
> 
> 
> On Sat, Jul 18, 2009 at 11:24 AM, AUDITORY automatic digest
> system<LISTSERV@xxxxxxxxxxxxxxx> wrote:
> > AUDITORY Digest - 17 Jul 2009 to 18 Jul 2009 - Special issue (#2009-163)
> >
> > Table of contents:
> >
> > AUDITORY Digest - 16 Jul 2009 to 17 Jul 2009 (#2009-162) (3)
> >
> > AUDITORY Digest - 16 Jul 2009 to 17 Jul 2009 (#2009-162)
> >
> > Re: AUDITORY Digest - 16 Jul 2009 to 17 Jul 2009 (#2009-162) (07/18)
> > From: Jont Allen <jontalle@xxxxxxxx>
> > Re: AUDITORY Digest - 16 Jul 2009 to 17 Jul 2009 (#2009-162) (07/18)
> > From: Christine Rankovic <rankovic@xxxxxxxxxxxxxxxx>
> > Re: AUDITORY Digest - 16 Jul 2009 to 17 Jul 2009 (#2009-162) (07/18)
> > From: Jont Allen <jontalle@xxxxxxxx>
> >
> > ________________________________
> > Browse the AUDITORY online archives.

-- 
                                                                       
 ,dPYb,,dPYb,            I8                                            
 IP'`YbIP'`Yb            I8        http://www.flatmaxstudios.com/
 I8  8II8  8I         88888888     http://www.flatmax.org
 I8  8'I8  8'            I8                                            
 I8 dP I8 dP   ,gggg,gg  I8   ,ggg,,ggg,,ggg,    ,gggg,gg    ,gg,   ,gg
 I8dP  I8dP   dP"  "Y8I  I8  ,8" "8P" "8P" "8,  dP"  "Y8I   d8""8b,dP"
 I8P   I8P   i8'    ,8I ,I8, I8   8I   8I   8I i8'    ,8I  dP   ,88"  
,d8b,_,d8b,_,d8,   ,d8b,d88b,dP   8I   8I   Yb,d8,   ,d8b,dP  ,dP"Y8, 
PI8"888P'"Y8P"Y8888P"`Y8P""Y8P'   8I   8I   `YP"Y8888P"`Y8"  dP"   "Y8
 I8 `8,                                                                
 I8  `8,      
 I8   8I  Public Projects :
 I8   8I  http://sourceforge.net/search/?type_of_search=soft&words=mffm
 I8, ,8'  http://www.psysound.org
  "Y8P'