Skip to content
HearingTracker
Log in

Can anybody explain technical differences between Phonak's AI in Sphere and Oticon Intent and Starkey Edge AI?

This topic is closed.

57 replies3k viewsLast reply

Jump
Link to this topic

In reply to #23

You didn’t understand what I meant. This is just an assumption, I don’t know the exact specifications. Let’s say the Phonak processor is made in 8 nanometers and the Starkey in 5 nanometers. The Starkey can cram more transistors onto the same board and not only does it have a more powerful processor, it will also consume less power. I think that’s why Starkey managed to increase the autonomy to 51 hours. And they said they had a guy from Intel who helped them develop it.

Sphere is new, but they used an older chip that we know from Lumity and added a chip (deep sound). and some other additions, like a new bluetooth module. What I mean by that is, they didn’t start from scratch like Starkey did with the Genesis version. … I don’t mean to say that Phonak is bad, I just want to point out why Starkey is 2 steps ahead.

It is possible that with the next version, Phonak will present a new chip that they designed from scratch. Manufacturing a chip from scratch is a long and slow process. Starkey also said that they worked on the chip for 5 years. I expect that Starkey will use that chip for at least another 5 years and make minor changes to it.

Edited

According to this:

Connectivity: Ultra-responsive ERA™ chip at the heart of Infinio | Audiology...6

Outperforming the competition and ready for the future, ERA connects your clients to their digital universe like never before. Learn more in this article.

The ERA chip is new and a successor to the Lumity chip.

Other literature states that the Deepsonic chip is new as well.

Btw. You are saying that Phonak might design the next chip from scratch, but that’s what they did with the deepsonic chip. If you listen to their product development manager, the started development I think in 2019. I don’t know how long the chip itself took, but it’s proprietary, meaning it was built from the ground up - like the Starkey chip. Yes, I agree that the Starkey Edge hearing aid was a complete redesign - the software platform as well - so who knows - they may have a superior product?

So do you agree that Starkey is 2 steps ahead? Currently Phonak can’t offer 51 hours of use or a full day of use with the Sphere.
I would love to see a review with a detailed analysis of how good the Edge mode is compared to the Sphere. This could ultimately help the buyer decide for themselves which is better for them, and whether they would prefer a longer battery life.

Just a parallel case.
the latest Pixel phones use an additional chip for AI and photo processing and are ahead of the bunch in that respect. I have no proof, but I think they have shared some of that with Samsung for their S24. Pixel use a few components from Samsung (notably antennas) so it could be a reciprocal deal.

In reply to @markoo355

Quite honestly, I would say that it depends on what the criteria is. Even then, it’s difficult to answer objectively.

If battery life is the main criteria, then clearly the Starkey has the longer life.

Speech in noise? Well - we have to wait for user reports, particularly those who have trialled both.

Then there is speech in quiet? Particularly group settings. This is overlooked in my view and it’s crucial for severe/profound loss. They say cochlear implants are superior in this area.

Then there is bluetooth, reliability and factors like ease of use and the app. I think you can’t declare one outright winner easily as we all have different priorities depending on our loss/lifestyle etc.

In reply to @markoo355

You really seem inclined to favor the Starkey. The only objective information I’m aware of is Hearing Tracker’s comparison scores and Phonak wins on that basis. I’m really more interested in technical differences and I have the feeling that info is only in the minds of people who aren’t sharing it. My suspicion is that Phonak is doing something different with it’s than either Starkey or Oticon since it seems to be throwing more computing power at it. Whether this results in any better results is anybody’s guess. If Starkey (or Oticon) is achieving comparable results, my suspicion is that it is a result of better software rather than hardware.

Edited

In reply to @glucas

If you search the site I have reported on my experience with the Starkey.
I have no experience with Sphere.
I suspect they are comparable with just different approaches.
As far as quite, noise wherever.
I hear well in all environments.
I just used edge mode practically all Thanksgiving.
I could hear and hold conversations.
I used to just mute my old aids.
I’m sure the a Phonaks do the same thing.

I have severe distortion in my left ear.
I always had problems in high nosie environments.
With the Starkey’s I can hear.
I may not have the-15 DB loss but my distortion really sucks.

In reply to @MDB

@MDB said:

My suspicion is that Phonak is doing something different

I think so too.

It has been getting harder and harder to know exactly what hearing aids are doing over time, which is frustrating for those of us who really do want to know. Adding is not going to make this easier as it gets pretty black-boxy. I’ve been chatting with a friend who is a bit closer to the /computer industry over in california and from what I gather right now the size and the battery draw of the Phonak device are actually meaningful (think of the power draw of, say, chatGPT). The deepsonic chip is custom built to run that particular DNN which is probably a larger DNN than the others, which does matter. There is distillation work that you can do after you develop the DNN to make it smaller, but it’s unlikely that the hearing aid manufacturers have reached that step yet.

That said, if someone does well in noise situations already it’s unlikely that moving to the Sphere will be that meaingful to them. Although our clinical tools for measuring speech-in-noise deficit are pretty gross right now, so one is only left with a pretty rough idea about how well someone will do in noise and it’s hard to say “this is actually over-powered for you” rather than “let’s not leave anything on the table in case we really need it in the next five years.”

Additionally, if anyone else was playing Final Fantasy 7 back in the day on the PS2 while the XBox was coming out, “new” technology does not always beat out a fantastic implementation of older technology.

It will be interesting to see where things go in the future with selecting brands of hearing aids when this DNN stuff has progressed a bit. But I think we also have to remember that we are talking about pretty razor thin differences at this point. This forum specifically attracts the sorts of people who get a bit obsessed about these things, but an average user who walks into a clinic and walks out with ANY of these hearing aids is going to be fine.

@Neville said:

Although our clinical tools for measuring speech-in-noise deficit are pretty gross right now, so one is only left with a pretty rough idea about how well someone will do in noise and it’s hard to say “this is actually over-powered for you” rather than “let’s not leave anything on the table in case we really need it in the next five years.”

I wonder what importance ACT (Audible Contrast Threshold) would achieve in assessing necessary noise reduction power. However, I think it’s not validated against Spheric Speech Clarity.

@Neville said:

Additionally, if anyone else was playing Final Fantasy 7 back in the day on the PS2 while the XBox was coming out, “new” technology does not always beat out a fantastic implementation of older technology.

I appreciate the comparison :smiley:

@Neville said:

This forum specifically attracts the sorts of people who get a bit obsessed about these things, but an average user who walks into a clinic and walks out with ANY of these hearing aids is going to be fine.

Good point. Forum users are often interested in their hearing loss. Many people outside the forum do not even have their in their homes, and their knowledge about hearing is very limited.

@Bimodal_user said:

I wonder what importance ACT (Audible Contrast Threshold) would achieve in assessing necessary noise reduction power.

The ACT is a demant product and they are currently only making it available to clinicians who purchase their updated equipment ($$ $$$), so I haven’t been able to try it out. (Honestly, this confounds me, the ACT does not require new equipment and should be easy enough to make more widely available.) However, it’s not really any finer of a tool than the as far as I can tell.

Edit: But you could probably roughly say that a score of -4 to 6 on the ACT probably doesn’t need the sphere program to do okay in noise.

Edited

I have never been involved in a speech-in-noise test such as or , so my opinion is only a supposition.

However, I think that when we discuss bells and whistles like , we shouldn’t forget about basics like possibly more closed acoustic coupling + …

Recently, I had my first opportunity to try the automatic “ordinary” Speech in Noise feature in my Paradise 90-RT with an earmold with a 1 mm vent (compared to 2,5mm diameter until recently) and increased gain in 1,5- 3 kHz.
I did not need to use my Roger On microphone like 1 year ago in the same restaurant with the same people.

@Bimodal_user said:

I think that when we discuss bells and whistles like DNNs, we shouldn’t forget about basics like possibly more closed acoustic coupling + REM…

Absolutely.

On that note, I wouldn’t go out of my way to fit a Sphere on a user who is definitely going to be an .

In reply to @Neville

@Neville said:

an average user who walks into a clinic and walks out with ANY of these hearing aids is going to be fine.

Thanks Neville.

One of my issues is speech in noise. I attended a session in an old school Lunchroom…table of 4 with 15 of us. Box lunches. No alcohol. I couldn’t understand anything my group of 4 said. Babble all around. Pulled out my iPhone. Tried speech in noise. Speech in loud noise. Mask. And more. Mask was best but still useless. Another issue is accents of any sort.

Paradise P90R’s. power receivers. New after warranty replacement. Just had . used P90’s now for 3 years.

Edited

You’re in those for another two years. I don’t know whether there’s something twitchy in your fit, or whether that asymmetry in your loss kills your speech in noise.

If you write up a letter detailing the sorts of group situations you’d like to be able to hear in and why a one-to-one partner mic won’t work, your clinician can submit it to WSIB to request a Roger On. Personal requests carry some weight with WSIB. With the new program, you’ll have to wait three months and one day since your last service. And you’ll probably need an updated test.

Edited

Thanks Neville.
Much appreciated.
Dave
Mississauga.

Edit…you and your colleagues here make a huge difference for me

In reply to @DaveL

I trialed the Intents and Sphere’s simultaneously for two months (generous audiologist). The Sphere’s definitely require more interaction with the app and the app is more complicated, BUT, therefore more feature rich. For me, the Sphere’s were much better in noise than the Intents and streaming with iOS devices, which is important to me, but not to many others…

Related discussions