Skip to content
HearingTracker
Log in

What's with the crazy hearing aid pricing differences in Costco versus chain hearing clinics?

This topic is closed.

221 replies14.3k viewsLast reply

Jump
Link to this topic
Edited
@1Bluejay said:

NOT to diminish Costco for one second, but I do not feel confident that my local store would have a dedicated audi as competent, patient and accessible as my current audi.

THAT’s a pretty good contribution to this discussion. In fact, as I read all of the above content in this thread, it looks to me that the success, or partial success, or failure, of an to meet a specific user’s needs will depend upon the COMBINED quality of at least all of the following separate components:

  1. The HA physical implementation
  2. The reliability of that physical implementation
  3. The software provided IN the HA
  4. The EXternal software provided that enables the user to make choices and changes to the HA
  5. The financial ability or willingness of the user to pay
  6. The wisdom of the in-canal choices made (Open, partially open, tulip, mould, vented or non-vented, M& or not, etc) for this specific user’s needs and wants
  7. The wisdom of the software tweaks made to cusotmize to the user
  8. The degree of support available for iterative improvements (time AND skill)
    9.The quality of HA maintenance
  9. The degree of persistence of the user to ensure all of the above
  10. The degree of persistence of the audi to ensure all of the above

Some of these are independent of the audi.

The remainder though are made or broken by the audi.

The problem though is that while you can find and read reviews of the different HA models available to buy from various sources in various selling channels, you cannot find and read reviews on the audis that are realistically available to choose from in your location.

If you have not previously used an HA, or if you relocated to a new location, you don’t know whether the audi you happen to encounter is experienced or inexperienced, detail oriented or not, perfectionistic or not, and persistent or not persistent.

Sure, you can “try” an HA that you have reliably heard is “a good one” and is apparently best suited to your needs, wants, and budget, but if the audi you work with can’t get you want you need and want, do you exercise the return privilege, perhaps repeatedly, going to a different seller location to ensure you can get a different audi, until you get an audi that gets it right for YOU? And, how do you KNOW that he or she got it as good as it can be, if this is your first pair of HAs, so that you have no comparative basis for comparison?

And, how do you know that a chain clinic with higher prices is going to have better audis than a Costco? Do chain clinics pay their audis better to get better ones?

Is it considered rude to ask the audi who first tries to work with you how much experience they actually have? Is it somehow possible to get references on a specific audi?

If the process is anyhting like finding a good car mechanic, or a good plumber, or a good doctor, I fear that the process is not easy in real life. I suspect it might end up being simply “the luck of the draw”. Am I wrong about that?

Jim G

Edited

In reply to @mami2529

@mami2529 said:

So… read this far now. It safe to assume Philips 9040 is as close (not 100% so) as Oticon’s Real?

I surprise Philips seems to get the “most of” technology from Oticon/Demant while Bernafon/Sonic getting less or almost outright different inside. e.g. No AI use in Bernafon Alpha XT.

If I want Oticon Real but I want it cheaper and don’t mind the hearing aids shell will not looks the same. So Philips 9040 is it?

I would say that the key difference between the Oticon Real and the Philips 9040 is in their brains. Yeah, both brains are powered, but their brains have different goals, are differently structured, and are trained differently. Other than that, all the body and arms and legs and head and eyes and ears are pretty much the same.

The Philips AI’s brain main goal is to use AI to clean up the speech diffused by the noise. They take hundreds of thousands of purely clean speech input samples (to be used as the reference speech-alone clean speech) and diffuse them (“dirty” them up) with noise. Then they feed each sample into an AI processing module that’s designed to remove the dirty noise from the speech. The cleaned up output is then compared to the referenced output (the purely clean sample they originally created). If they don’t match well, the discrepancy is fed back into the AI so that the AI can be tweaked to minimize this discrepancy and does the cleaning job better next time on another sample. They train the AI using hundreds of thousands of speech samples until the AI is good enough. Below is a screenshot of the flow.

How is this different from the traditional frontal beam forming directional approach? Well, this approach blocks the surrounding sounds except the front where speech is, but there still remains the issue that the front speech is still dirtily diffused with the same noise. What the user ends up hearing is the speech that is still laced/diffused with the dirty noise. The user would still have to filter out this diffused noise using their brain. That may be easy or difficult depending on how bad this diffused noise is that is laced onto the speech. The Philips AI goes one step further and cleans this dirty boy up, too.

Another thing is that with the cleaner speech coming out of the Philips AI, it’s easier to recognized and be more easily understood, so it really stands out clearly and even if some of the surrounding noise is maintained and tolerated and not blocked out by directional beam forming, people can still zero in and single out this speech amongst the noise.

image

The AI brain of the Oticon Real (and More) is structured and trained differently than the Philips AI. It’s not designed and trained to clean up noisy speech samples like how the Philips AI brain is. Instead, it’s designed to take in a WHOLE sound scene (not just a meager noisy speech input at a time, like with the Philips AI) and uses a deep neural network () to basically break down the components of this sound scene, propagate them through the DNN and effectively rebuild the whole sound scene again at the output. Then this “virtual” sound scene is compared against the real (actually captured) sound scene to see how they are different. The discrepancies are fed back into the DNN to be tweaked in a way that those discrepancies are minimized in the future.

In the example screenshot below, you can see an example sound scene whose components are the sound of the seagull, the speech between the woman and man, and the music from the tuba player, and probably more, not just limited to those 3 components. The DNN breaks down the components then recombines them in the end to “rebuild” or “recreate” the sound scene.

The Oticon AI DNN is trained with 12 million sound scenes (compared to hundreds of thousands noisy speech samples for the Philips AI) until Oticon deemed it good enough to release into production. As you can see, the Oticon AI/DNN brain and the Philips AI brain are designed to manage sounds in completely different ways. The Philips AI is focused SOLELY on removing diffused noise from noisy speech samples. The Oticon AI is focused at a much broader level, not just speech clean up, but at breaking up ALL components of a sound scene and rebuilding them again. Obviously the Oticon AI/DNN attacks the sound processing issue at a much bigger macro level (the whole sound scene), while the Philips AI attacks the sound processing issue at a much smaller micro level (just cleaning up the noisy speech).

The next logical question is what good does it do for the Oticon AI/DNN to break up a whole sound scene into components only to then get rebuilt back into a whole sound scene again? The answer is that if the Oticon AI/DNN can break up and rebuilt a sound scene accurately, then it can be made to manipulate the components of that sound scene to give a more balanced scenario in however way you want it to be.

In the example below, if the sound from the tuba player is too loud and overwhelms the conversation the man and woman are having, then if you set your Neural Noise Suppression to a high level (aggressive suppression), then the tuba player sound component can be reduced as to not overwhelm the speech between the man and woman. If the Neural Noise Suppression setting is set to low, then maybe the tuba player sound component would not be suppressed as much, and maybe the sound component from the seagull might be enhanced a little bit more, all the while everything still gives deference to speech because speech always has high priority. Overall, the ability to break down and rebuild all sound components of a sound scene accurately will open the way for the AI to be able to rebalance a sound scene to the way that YOU can control and tell the AI how to do it, via the settings.

image

Edited

In reply to @Neville

@Volusiano said:

You’re supposed to get speech detection 360 around you with the open paradigm in the Real.

@Neville said:

When the Opns came out, the way they were creating the “open paradigm” was to check for speech right behind and if it wasn’t there take a ‘snapshot’ of the noise behind and delete that from everything in front under the assumption that what was behind, as long as it wasn’t speech, was an average sort of unwanted noise that it could then comb out of the front to better focus on wanted speech sounds. Clever way to reduce background babble sound. However, if what was behind you was near speech they wouldn’t want to use that to substract from the front as it might subtract more desired signal. The solution was that if the Opn detected speech close behind you, the “open paradigm” function would temporarily turn off. So speech right behind was specifically NOT something that the new “open paradigm” was helping with, and the opn settings were essentially a fairly wide directional setting but with that extra processing to reduce unwanted noise within that directional bubble, in contrast to other devices’ narrow beam forming strategies.

With the More, unwanted noise changed from “non-speech-stuff-in-back” to “what-this-AI-says-is-noise”. Also cool. But as hearing aids advance, they still hold onto the developments that have been made over time. It’s getting harder to know exactly what is going on under the hood, but elements of the Opn may remain and one of them may be a fairly consistent null in back. I’m not sure. But that would be different than 360 degree adaptive directional beam forming.

What @Neville said above is right on the dot. I just want to add a couple more charts to help support and illustrate what he said even more clearly.

Below is the flow chart for the original Oticon OPN and OPN S, which is called the OpenSound Navigator (OSN). The Balance module seen in the middle does the MVDR beam forming to create the null directions, as illustrated in the second screenshot. But as @Neville explained, there’s a Voice Activity Detector in the OSN such that if speech is detected in the back or front, the Balance and Noise Removal modules would be frozen to let you hear that rear speech. But now you don’t get the benefit of having any kind of noise reduction/removal anymore because the OSN is frozen until rear speech stops.

image
image

Starting with the More with the , this OpenSound Navigator (OSN) technology is still kept and utilized, as seen in the screenshot below. The OPN Analyze module above is now called the More Scan & Analyze module, the OPN Balance module above is now called the Spatial Balancer, and the OPN Noise Removal module is now replaced by the DNN Neural Noise Suppression module.

image

But the More flow no longer has a Voice Activity Detector in place like the OPN/S OSN had, which was used to freeze the Balancer MVDR beam forming and freeze the Noise Removal, as to let this rear voice through. That is because the DNN does the job of the voice activity detection indirectly now. The DNN just picks up all sound components from all around (voices included) and process them in its own way to rebalance them.

As to why @Markem is not able to detect speech behind him on the Real, even without a Voice Activity Detection like the OPN had, the Real DNN should have picked up this rear voice component and process it to be heard. The most obvious thing I can think of that can prevent this is if the Directionality Settings in the Real Intelligence menu page is not set to Neural Automatic like it’s supposed to, but is set to Full Directional, which implies front beam forming only.

If it IS set to Neural Automatic, then the second guess I have is that maybe the Neural Noise Suppression is set to be too aggressive, and if the rear speech is not picked up cleanly and clearly but is more of a babble, then babble is considered as noise and not clear voice speech. But this is just a guess and who knows for sure. It may be a question for an Oticon rep to answer if the is also stumped.

Thank you @Volusiano for a very detailed and informative explanation.
I would absolutely love to hook up the Reals to the Noah/Genie to solve the mystery of the microphone settings.
My fear at this point in time, without any back up ’s, I’m hesitant to hook up the Reals to the Noah/Genie. I have this fear that somehow I’ll mess this up. I may have to be patient a tad longer to play around at least until I have the 9040 backups in my possession.
Who knows, the way you explained the inner workings of both it may be possible that the Reals end up being the backups. Wouldn’t that be something considering the extreme cost differences.

I’ve got a highly technical background. I spent nearly 30 years in computer development at IBM and for better than a year had an early effort going on in the fiefdom that I managed. And I don’t care about any of this technical stuff w/respect to Oticon versus Philips except out of general curiosity. I just want to hear better. I am happy to pass along my experience of the 9040 versus all of my previous , the most recent having been the KS10 FWIW.

The way that I would characterize the Philips versus KS10 “openness” can be illustrated with what happens when I walk into a big store…, say, a Costco store, lol. With the KS10 I would walk in and pretty quickly my HA would noticeably change. All of the ambient noise would get muffled, as though I had cotton in my ears. Voices in front of me would be good. While I understood why this was happening - trying to help me with speech-in-noise - this effect annoyed me and I never go used to it. I wanted to hear what was going on around me.

When I read on this forum about the way that Oticon works with the open sound landscape, I was interested and willing to consider the huge price difference. Before I could line up an Oticon trial I learned about the 9040 and tried it and liked it so well that I dropped my pursuit of Oticon due to the huge price difference (as well as my positive past experiences and comfort w/Costco HA centers). With the Philips this sort of shutting-down-the-environment effect never happens. I hear all of the sounds inside a Costco. I can sense NO attenuation of sounds either with or without my wife or others talking to me.

I have experimented with my wife. I listen to the sounds of the store. Then I have her talk, not listening to what she is saying. Rather I continue to focus on the background noise, to see what the HA are doing with it. I sense nothing different about the background noise, whether or not she is talking. What I do sense is tha the HA are making my wife’s voice sharper; they sort of pick her voice out of the humdrum and make it better. This to me is an open sound landscape, and I really, really like it. Yes it can be annoying at times, but on balance this is so much better for me. It is a more natural sound environment, like I remember from the days when my hearing was good.

(I have the Roger Select for my KS10, a great device. It is the best thing ever, for restaurants. But I could never remember to bring it w/me. The 9040s are better for me, overall, because they are always with me. 9040 aren’t as good with voices, but it’s been at least passible in the loudest restaurants, and much better than the KS10 when I forgot my Roger Select. I don’t have a 9040 TV streamer. I’ve kept my KS10 which have a streamer via Roger Select. Nothing beats TV through a streamer, so occasionally I have to set aside the 9040 and get out the KS10/Roger. One of these days I’ll break down and get a streamer for the 9040.)

I don’t understand all of the debate and rationalization of sameness/difference between Oticon and Philips. They are cousins. They are different, but they rhythm. None of the technical comparisons mean a thing to me. My lived experience is what I care about. And I would say that if you are a person who prefers to hear all of the sounds all of the time, as with me, and you don’t have that experience with your present HA, and you don’t like Oticon prices, you might want to give the 9040 a shot and judge it on its own merits vis a vis what you have, and not worry about whether or not it is an Oticon. There is little to lose, given the Costco 180 day thing.

Edited

Exactly why I returned the Phonaks. Love hate relationship with those. Too much hate, not enough love.
The app is another story. Real app is very glitchy, hopefully the 9040 works better.

@Torbill said:

The way that I would characterize the Philips versus KS10 “openness” can be illustrated with what happens when I walk into a big store…, say, a Costco store, lol. With the KS10 I would walk in and pretty quickly my HA would noticeably change. All of the ambient noise would get muffled, as though I had cotton in my ears.

In reply to @Torbill

@Torbill said:

You might want to give the 9040 a shot and judge it on its own merits vis a vis what you have, and not worry about whether or not it is an Oticon.

Amen to this! I currently wear the Oticon OPN. I don’t have the More or Real. Despite having a technical understanding of how the Oticon Ai works and the Philips works, and how they differ, when my OPNs bite the dust, my most likely choice of aid will be the Philips 9040 because it seems to deliver the best value for the money. And I won’t have any need to justify that it’s the same or almost the same as the Oticon aid in order to choose it.

The exceptional value it gives me at the $1500 price point is good enough already. Why would I want to be greedy by wanting to think that it’s the same as an aid that is 3 or 4 times the cost so that I may feel like I get even a better deal than I already get?

In reply to @Torbill

I have been reading your experiences with interest as I currently have KS10s, and haven’t been looking to change soon. I may reconsider.
From reading the Costco website, it appears the 9040s need an “audio clip” to “make hands-free phone calls with an Android phone”. Is make the same as receive, as in initiate vs answer? I am very fond of the hands-free ability of the KS10s to receive calls. If I have to add more hardware to get the same with the 9040s, that’s a deal breaker
Thanks…

@ Volusiano Good morning.
Directional settings for General 1 are Nueral Automatic. That looks okay there. Neural noise suppression I’ve uploaded a snapshot for review.

Neural

Not sure if this is relative to the General 1 program. When I dig a little deeper I am seeing setting as Fixed omni. Still new to the game, I’m not sure if the 2 are related.

REM

In reply to @oetbyg

@oetbyg said:

I have been reading your experiences with interest as I currently have KS10s, and haven’t been looking to change soon. I may reconsider.
From reading the Costco website, it appears the 9040s need an “audio clip” to “make hands-free phone calls with an Android phone”. Is make the same as receive, as in initiate vs answer? I am very fond of the hands-free ability of the KS10s to receive calls. If I have to add more hardware to get the same with the 9040s, that’s a deal breaker
Thanks…

I think the Philips 9030 and 9040 has support that in theory should work directly with Android phones that also has ASHA support. In practice, this is a new territory so whether they work well together probably depends on the phone you have.

But if your Android phone is an older model that doesn’t have ASHA support, then yes, you’d need to use the Philips Audio Clip to work with it.

Edited

In reply to @Markem

@Markem said:

@ Volusiano Good morning.
Directional settings for General 1 are Nueral Automatic. That looks okay there. Neural noise suppression I’ve uploaded a snapshot for review.

Neural

Hi @Markem , I don’t see anything out of the ordinary here that would prevent you from hearing voices coming from behind with these settings. The main thing is to make sure it’s not set to Full Directional, and it’s not.

I would suggest maybe you try some controlled experiment with the help of somebody to be a speaker and they rotate their positions around you to see if you can still have a noticeable difference between the various positions.

I don’t know the circumstance(s) you were in when you noticed that you couldn’t pick up the voices from behind you. It might have simply been that there were a lot of things going on and the distractions took your brain hearing away from the voice in the back so as to make you not notice it as much. What I’m saying is whether it’s possible that the voice was there but your brain didn’t notice it for some reason, therefore leading you to conclude that it wasn’t there but it was maybe because you didn’t notice it? That’s why a more controlled experiment would help make your observation more conclusive one way or another.

Even without controlled experiments, next time you’re in a restaurant and you still can’t pickup conversations of people in a table behind you even when you’re deliberately paying attention, then a more definitive conclusion like that can be brought to your audi to ask Oticon support why and get some further guidance from them.

I also notice that you have a pretty hefty hearing loss across the board based on your . I can’t help but wonder if this somehow affects the way the Oticon works on how the rear voice gets presented to you.

Another easy thing to try is be in a program like MyMusic or whatever that has the Directionality Settings set to Fixed Omni, and do your controlled experiment to see if Fixed Omni even let you pick up rear voices or not. This is just a sanity check to see if anything else is at play, because Fixed Omni should pick up everything you can hear around you by design.

In reply to @Markem

@Markem said:

Not sure if this is relative to the General 1 program. When I dig a little deeper I am seeing REM setting as Fixed omni. Still new to the game, I’m not sure if the 2 are related.

REM

I can’t really comment on this because I don’t know how your audi did for you. If REM was done by using Genie 2 and its REM Autofit along with a compatible REM equipment brand, then maybe this has some meaning. But if REM was done outside of the Genie 2 REM Autofit using an external REM equipment, then what you see here may not mean much anyway.

But even if there is some meaning for this, the fact that it’s set to Fixed Omni, if it carries over to what you hear in your default P1 program, it should actually let you hear 360 degrees around you and be able to pick up a rear voice, because that’s what Fixed Omni implies.

In reply to @Volusiano

@Volusiano said:

what good does it do for the Oticon AI/DNN to break up a whole sound scene into components only to then get rebuilt back into a whole sound scene again? […] it can be made to manipulate the components of that sound scene to give a more balanced scenario in however way you want it to be.

In the [Oticon DNN diagram], if the sound from the tuba player is too loud and overwhelms the conversation the man and woman are having, then if you set your Neural Noise Suppression to a high dB level (aggressive suppression), then the tuba player sound component can be reduced as to not overwhelm the speech between the man and woman. If the Neural Noise Suppression setting is set to low, then maybe the tuba player sound component would not be suppressed as much, and maybe the sound component from the seagull might be enhanced a little bit more, all the while everything still gives deference to speech because speech always has high priority. Overall, the ability to break down and rebuild all sound components of a sound scene accurately will open the way for the AI to be able to rebalance a sound scene to the way that YOU can control and tell the AI how to do it, via the settings.
[emphasis added]

I am perplexed as to how the wearer can control the Neural Noise Suppression balancing “via the settings” unless that means via the Genie 2 fitting software as operated by a provider or . The only controls/settings available to the Oticon Real wearer are in the Companion smartphone app that provides no Neural Noise Suppression control. It supplies a “SpeechBooster” on/off control that works by “focusing on speech sounds in front of you and suppressing other sounds around you”; it’s a directionality control.

The settings provided in Genie 2 fitting software that might relate to Neural Noise Suppression by the are:
– Neural Noise Suppression: on/off toggle;
– Neural Noise Suppression: a 0/2/4 control for “easy” situations and a 6/8/10 dB control for “difficult” situations, for degree of suppression;
– Directionality settings: a menu with choices of Neural Automatic, Full [forward] Directionality, and [open] Fixed Omni[directional]. Neural Automatic “adapts to all directionality settings while remaining as open as possible.”

An Oticon blog page about the More2:

We trained a DNN with 12 million complex real-life sound scenes like these, which it then learned to analyze, organize and balance. After it had learned all this awesome knowledge, it was ready to power Oticon More. Now this hearing device can utilize the DNN’s intelligent capabilities when balancing and prioritizing the sounds that are important to you. [emphasis added]

(Yeah, it really does say “awesome knowledge”. I’m sure the engineers who worked on the DNN love the marketing blather.) Maybe the DNN can distinguish types of non-speech sounds that are potentially valued, e.g., tubas and seagulls, from never-wanted noise. But given that the fitting software, much less the Companion app, has no “tubas and seagulls” setting, the best that the DNN can do is suppress some non-speech sounds a bit less than others.

To whom will this setting make a difference? I’d think that would be only wearers with moderate to mildly severe losses who are generally satisfied with their speech understanding in scenarios containing competing but wanted sounds.

In reply to @Volusiano

Thankyou. This is something I will bring up with my Audi. After having a chance to discuss this here that will add value to my next hearing appointment I’m sure. And for that I’m grateful for this website and the knowledgeable members within.

In reply to @JimGnitecki

OW! You have an encyclopedic understanding of the issues here, JimGnitecki! You should be a columnist somewhere writing on our quest for the perfect solution! :slight_smile:

As you summarized, it is multi-factorial. To niggle, I’d add: audi’s knowledge of ACCESSORIES is critical! If they don’t know it, they better have a hot line in to Phonak, Oticon, or wherever to get the answer. My audi has a dedicated Phonak rep right in her geographic area, and I got to meet that rep on 2 occasions prior to buying the Phonak accessory Roger ON iN V2 “stick” for streaming TV and other uses.

Yes! It was a stressful challenge for me to find this audi after having the perfect guy for 15 years. But we moved cross-country, and first order of business was to get in touch with an audi. I searched online at HealthGrades, et al, but it was not perfect. Many reviews are leaving Yelp for Google, so that didn’t work either. Ultimately, I found enough info on the new audi’s education and background, and while she’s a relative “newcomer” to the profession, by golly she is PERSISTENT!! So I lucked out.

YES by all means I’d try out different audis AND different model/makes till I MYSELF was satisfied. These are not cheap! When we test drive cars, we go to a few dealerships, maybe weigh the advantages, cost and service options of one place over the other before we pull the trigger. It is your right to be particular and choosy! And you’re also right in observing that it’s as challenging as finding a reputable, competent “service” professional (plumbing, garage, electrician, etc.,)

I may have a slight advantage nailing down an audi over others cuz I am HIGHLY articulate, experienced and nuanced in my hearing wants and needs. :grinning: I would never find myself talked into a make/model that doesn’t work for me, and if the audi is unwilling to view me as a partner for life, I’d look elsewhere. I once had an audi tell me, “My goal is to get you set up today and NEVER SEE YOU AGAIN.” I blanched, staggered out the door - kept those aids - but never ever went back. I found a new audi who took over where that inflexible gal left off.

Luck of the draw. Experience. Location. Word-of-mouth. Articulation. All these come in to play!

Edited

In reply to @brec

@Volusiano said:

what good does it do for the Oticon AI/DNN to break up a whole sound scene into components only to then get rebuilt back into a whole sound scene again? […] it can be made to manipulate the components of that sound scene to give a more balanced scenario in however way you want it to be.

@brec said:

I am perplexed as to how the wearer can control the Neural Noise Suppression balancing “via the settings” unless that means via the Genie 2 fitting software as operated by a provider or DIY. The only controls/settings available to the Oticon Real wearer are in the Companion smartphone app that provides no Neural Noise Suppression control. It supplies a “SpeechBooster” on/off control that works by “focusing on speech sounds in front of you and suppressing other sounds around you”; it’s a directionality control.

The settings provided in Genie 2 fitting software that might relate to Neural Noise Suppression by the DNN are:
– Neural Noise Suppression: on/off toggle;
– Neural Noise Suppression: a 0/2/4 dB control for “easy” situations and a 6/8/10 dB control for “difficult” situations, for degree of suppression;
– Directionality settings: a menu with choices of Neural Automatic, Full [forward] Directionality, and [open] Fixed Omni[directional]. Neural Automatic “adapts to all directionality settings while remaining as open as possible.”

When I wrote that post earlier (which you quoted here), I thought about that, too (along your line of questioning here). Naturally I should have clarified that the user-control aspect is not directly via the Companion or ON app to be set by the user, but only via Genie 2 by the or a . So thanks for pointing that out for clarification, too.

You are correct that you can only toggle the Neural Noise Suppression ON/OFF, and control its setting values for Easy and Difficult. And the Directionality Settings.

But there is another setting on that page → the Environmental Configuration to allow either experienced users or inexperienced users to define which of the 5 listening situations listed they deem to be easy and which are deemed to be difficult. That is definitely part of the Neural configuration as well.

There is another aspect in Genie 2 called the Fitting Assistant. It’s a helpful tool to help you get Genie 2 to be set up based on more laymen terms that may be more easily understandable for non-techies. But to be honest, I think that answers to the Fitting Assistant will simple lead to an appropriate setting combination in the MoreSound Intelligence page. So it’s not like the answers in Fitting Assistant can lead to even more controls that you wouldn’t already have in MoreSound Intelligence.

I personally think that the amount of control Oticon allows the user to have in the Genie 2 MoreSound Intelligence is good enough to give sufficient directions to the DNN to adapt its balancing act more to the users’ preferences. I think Oticon needs to walk a fine line between giving the users enough control, but not too much control as to present conflicting operations inside the DNN.

I’m sure that there are also many different and more abstract things used in a human brain neural network, like pattern recognition, elimination process, memorialization, deductive reasoning, etc. etc. that are inherent to the brain’s neural netowrk and get used by the brain hearing to process sounds. These things may be modelled and built into the Oticon DNN in such a way that the break down of the sound scene into sound components may help lend a hand in how to rebuild this sound scene up using these inherent processing “techniques” used by the brain. Of course it wouldn’t make sense to give users control of these abstract things, so that’s why I haven’t mentioned it until now. But apparently the Oticon whitepaper makes it very clear that their DNN flow approach is to take captured sound scenes and break them down into sound components inside the DNN and rebuild them up again.

In that detailed whitepaper (that came up as the first search item in your search as you mentioned in the other thread), in the Neural Clarity Processing (DNN) section starting on page 5, Oticon did brush over some examples of some of these abstract things in explaining why they think using a DNN would be a good approach to mimic the human brain in the processing of sounds.

In reply to @brec

@brec said:

An Oticon blog page about the More:

We trained a DNN with 12 million complex real-life sound scenes like these, which it then learned to analyze, organize and balance. After it had learned all this awesome knowledge, it was ready to power Oticon More. Now this hearing device can utilize the DNN’s intelligent capabilities when balancing and prioritizing the sounds that are important to you. [emphasis added]

(Yeah, it really does say “awesome knowledge”. I’m sure the engineers who worked on the DNN love the marketing blather.) Maybe the DNN can distinguish types of non-speech sounds that are potentially valued, e.g., tubas and seagulls, from never-wanted noise. But given that the fitting software, much less the Companion app, has no “tubas and seagulls” setting, the best that the DNN can do is suppress some non-speech sounds a bit less than others.

To whom will this setting make a difference? I’d think that would be only wearers with moderate to mildly severe losses who are generally satisfied with their speech understanding in scenarios containing competing but wanted sounds.

Haha, I fully agree that this Oticon blog is overreaching when it emphasizes the “prioritizing the sounds that are important to you” part. It’s simply not possible to go into that level of details and let the user specify something like “what’s important to me is music, human speech, nature sounds, and rank each type on a scale of 1 to 10”, for example. The only obvious thing a hearing aid can presume that’s important to the user is speech.

But I think the Environment Configuration of the 5 types of listening situations in the MoreSound Intelligence session is probably the most important control input that Oticon can give to the user. It allows the users to basically say I’m a novice user, or maybe one with heavier hearing loss, by setting Easy to include only Very Simple, and Difficult to be everything else. On the other extreme, it allows a very experienced user, especially one with a milder hearing loss, to say that Difficult for me is only Very Complex, and the rest (Very Simple, Simple, Moderate, and Complex) are actually Easy for me. And a moderately experienced user or one with moderate to severe hearing loss can choose appropriately between those 4 combinations.

I concur that whoever that writer who wrote that Oticon blog using word choices like “awesome knowledge” and “prioritizing the sounds that are important to you (with actual emphasis added)” did a very poor job and actually took away the credibility of the technology Oticon has instead of boosting it.

I pad $3500 for my pair of Phonaks a year ago.

The included two year protection / loss coverage has a $350 deductible.

I like the programming of my aids they sound great. But Phonak took several steps back with the design. I will not buy Phonak again in the future.

Ok. I spent £3,000 here in the UK for my Oticon Real 1’s. That is around $3670 US dollars including a 5 year warranty. I love these things but I have already had one pair stolen (!) and I had to claim on my insurance. I have a Noahlink btw and have been playing around with Oticon Genie. It is fascinating. I am going to buy a spare pair of ’s now as I travel frequently. I am off to Costco and will order myself a pair of Philips 9040’s if they are available. I will then report back on my experience and will try and get hold of the fitting software too.

Related discussions