Skip to content
HearingTracker
Log in

Adaptive Phonak Digital vs NAL-NL3

17 replies427 viewsLast reply

Link to this topic
Reply
Edited

I am experimenting with NAL-NL3 after having been a user and fan of Adaptive Phonak Digital for about 5 years. I wear Phonak Infinio 90-Sphere with cshells. I never liked the quality of singing voices when streaming music, or voices on televisions when using

The screen shots are (top 2) Calm Situation (algorithm in lower left) and (bottom 2) Spheric Speech in Loud Noise). Each algorithm was fitted with SoundRecover2 OFF and Semi-Linear Compression.

My initial impressions of NAL-NL3;

  • When I stepped outside I thought the crickets and cicadas were on steroids.
  • My footsteps on hardwood floor sound louder
  • Speech from TV, social media videos do sound crisper. Brass instruments during music playback sound crisper.
  • I could hear the crackle of a neighbors backyard fire as he burned yard waste 130 ft away.
  • Where I still struggle:
  • Retail point of sale where background is not high enough to activate Speech in Noise or Spheric Speech in Loud Noise
  • Elevators
  • Doctors exam room with doctor 12 feet away.
  • I don't seem to hear traffic in a parking lot as well with NAL-NL3 compared to A.P.D.
  • The Phonak startup jingle when I put the aids in at the start of the day is very low.

NAL-NL3 is supposed to have some built in module called "Comfort in Noise". I couldn't find an explanation in layman's terms as to how this activates or enhances the users experience.

I still struggle to understand why algorithms seem to put more emphasis on different parts of the spectrum. I thought the algorithms were supposed to match hearing aid output to hearing loss. Yes, I understand everything sounds different to every user. What may not be visible in the screenshots is, with respect to NAL-NL3, G80 gain exceeded Target G80 and G50 gain did not meet G50 Target gain. In looking at the charts, amplification starts to fall off around 2Khz. When fitted with APD, G80 gain also exceeded G80 Target gain and G50 gain was very close to G50 Target gain.

Someone once told me hearing aid algorithms are like driving a car from Chicago to New York. You can drive a Honda Civic or a Ferrari. They both will accomplish the same thing but one may be more comfortable than the other.

Cheers!

5 people in this conversation12 min readJump to latest

I’ve been experimenting with the same thing in Target 12. I used for a long time and tried NAL-NL3 because, on first listen, it did seem crisper and more detailed. But after living with it for a bit, speech in my inside and outdoors was actually worse for me, even though some sounds seemed more sharply defined. I ended up going back to APD.

That experience made me much less interested in whether one fitting curve appears to hit a target “better” on the screen. Two prescriptions can amplify the same very differently because they make different choices about gain, compression and how loudness is distributed across frequencies.

Your comment about retail counters especially caught my attention. I’ve also found that the difficult situations aren’t necessarily the really loud ones where Sphere kicks in — it’s often the awkward middle-ground environments where the automatic system hasn’t decided anything dramatic is happening yet.

So for me, NAL-NL3 gave more apparent detail, but APD gave me more usable speech. I’d be interested to know whether your opinion changes after several days rather than during the initial “wow, I’m hearing everything” period.

In reply to @orgullodemexico

Could you please specify which version of Phonak Adaptive Digital you're using? It is device specific, as in when updates are released to they are specific to the products they are launched with and not backwards compatible with previous product generations as is the case with NAL and formulae.

You can always adjust the speech in loud noise curve activation threshold or use a manual speech in loud noise program for the retail counter situation that you described.

@Hearing_Potential said:

Could you please specify which version of Phonak Adaptive Digital you're using?

I'm using Adaptive Phonak Digital 3.0 in my Infinios. I felt the same about ADP vs when I was using it in my Phonak Lumitys, which was version 2.0

For the quiet situations to which I referred, I have tried Restaurant program, Speech in Noise and Spheric Speech in Loud Noise. While background noise was reduced, so was the voice of the person speaking. I get tired of asking someone to speak up.

Edited

In reply to @orgullodemexico

@orgullodemexico said:

I thought the algorithms were supposed to match hearing aid output to hearing loss.

Depends on what you call "match".

What you said about hearing certain sounds match what your charts show. I'm only going to comment on the quiet situations one.

The sounds of crickets, cracking boards, increased clarity are high frequency sounds. In the charts (quiet):

Below 4Khz: About the same
4Khz: NAL-NL3 ~10%+ more
6Khz+: NAL-NL3: ~20%

This is what I expect to see when comparing a propietary formula vs NAL-NLX.

The elevators, rooms, traffic and noise sounds are mostly low to mid frequency sounds. Since the gains are about the same for both and NAL-NL3 your charts, for <4Khz, what you said matches the curves.

@orgullodemexico said:

I still struggle to understand why algorithms seem to put more emphasis on different parts of the spectrum

Because you don't need to increase gains at all frequencies equally to improve speech understanding. Speech sounds are not spread equally across all frequencies.

Different formulas have different approaches for speech improvement.
NAL-NLX focuses on overall loudness comfort and speech intelligibility (example: Preventing loud sounds from feeling piercing or overwhelming). This is done by using compression and setting gains according to what NAL determined is needed for speech.

DSLv5 is more linear. It's enhances speech by providing audibility of speech across frequencies but it does not have the same level of comfort as NAL-NLX.

Propietary formulas tend to put strong emphasis on comfort by prescribing just the necessary gain and increasing compression ratios (this can be particularly helpful for new users). This comment is extremely simplistic because different manufacturers have different approaches and different versions of propietary formulas. I've notice that the last APDs tend to prescribe higher gains than before.

Many people don't like high gains at high frequencies because it sounds uncomfortable, robotic, tiny , tiring, etc.

In short, the approach is different because its about winning the fight between "intelligibility and comfort" vs. "maximum audibility of speech".

@Hearing_Potential : Did I explain this reasonably well? Feel free to correct my mistakes.

In reply to @orgullodemexico

Is it with M receiver? Maximum possible gain seems low. Look at mine range, from M to UP. The shape are a bit odd because of + further adjustments:

UP in cShells with 1mm diameter vent.

Regarding my gains, these are DSLv5 Pediatric. I have chosen this, because it emphasises highest frequencies where I don't have dead regions and I am hard of hearing since 4 yo. And I like that DSLv5 take into account my from AudiogramDirect which I noted and used to adjust when I entered my "non-AudiogramDirect ".

I am surprised that despite your loss in the lows, you have almost no gain until 300-350 Hz in your right ear? I understand that there is prevention of upward spread of masking, but still - it's a bit strange.

I can't shake the feeling that "first-fit" in your screens is very conservative and delicate and you can get more benefits from your , especially if you haven't dead regions in high frequencies.

If I were you, assuming I want as much as possible from Infinio Sphere potential, I would at least try UP in cShell with no vent.

It's a pity that feedback threshold couldn't get more from your right ear 6-8 kHz - perhaps overtuning would help?

Edited

After I read your comment, I noticed more things:

1.- Target and prescription curves are "parallel": OP is likely self programming and has not done further adjustments beyond the prescribed values.

2.- I believe, in your case, you’re accustomed to seeing much higher gains because you use pediatric version ( I tested it once and I could not stand it for 1 min). And yes, DSL can prescribe higher values at high frequencies (which I prefer) but many find it uncomfortable. His gains are below target, not by much with but more with NAL-NLX. OP didn't say why.
3.-

@Bimodal_user said:

I am surprised that despite your loss in the lows, you have almost no gain until 300-350 Hz

Increasing lows may increase noise and not necessarily lead to better speech understanding (the fight between comfort and audibility that I mentioned earlier). While DSL increases gains across all frequencies (which is what you have), NAL-NLX considers comfort and that maybe the reason why the prescription is lower in the lows.

In reply to @orgullodemexico

@orgullodemexico said:

NAL-NL3 is supposed to have some built in module called "Comfort in Noise". I couldn't find an explanation in layman's terms as to how this activates or enhances the users experience.

How this works depends on how a specific manufacturer has implemented NAL-NL3. I haven't yet had any training on how Phonak has implemented NAL-NL3, but I would hope that they are implementing the NAL-NL3 comfort in noise module to prescribe gain in relevant destinations in Autosense OS when NAL-NL3 is being used as the prescription formula. The goal of the comfort in noise module is to provide a more comfortable listening experience in noisy situations while trying to maintain speech intelligibility.

One thing that I'm noticing is that people are using the gain curves in Phonak Target rather than the output view. You can get a much better idea of what hearing aids are doing by using the output view --> real ear . If you don't have the curve turned on in this view turn it on in settings.

In reply to @OcramSagev

@OcramSagev said:

@Hearing_Potential : Did I explain this reasonably well? Feel free to correct my mistakes.

Honestly, there are entire book chapters written about the differences between NAL and fitting formulae and the goals behind them. I don't have the time or energy to write about them, but I would refer you to Harvey Dillon's Hearing Aids: Volume 2 if you'd like to learn more. There's also this older article comparing NAL-NL2 and DSL6 with a short commentary from the leads of the teams involved in developing each.

In reply to @Bimodal_user

@Bimodal_user said:

Is it with M receiver?

M receiver

3 wire length

P power

I will see an audi at the end of this month for some speech in noise testing, tweaking, possible R.E.M. and discuss which algorithm would be a better choice for me.

Right now what frustrates me more than hearing conversation in a noisy background is those situations that seem to fall in between Calm Situation and Speech in Noise.

At home where it's 35db with nothing on, I can hear a hand held scanner that is 25 feet away. At a retail checkout, I can't hear the clerk that is 3 feet in front of me and it's below 60db. Yea, I tend to measure every venue I visit.

Edited

In reply to @OcramSagev

@OcramSagev said:

1.- Target and prescription curves are "parallel": OP is likely self programming and has not done further adjustments beyond the prescribed values.

Indeed, this is why I'm afraid that much potential benefits were left on the table. And I don't mean that OP should go with DSLv5. What I mean is making more adjustment than simply go with first fit, which is proven in many research papers insufficient especially in high frequencies.

@OcramSagev said:

2.- I believe, in your case, you’re accustomed to seeing much higher gains because you use DSL pediatric version

While you have grounds to think so, and I would think the same if I were you, that's not the case. I tried to go with 5.0 Pediatric ~2 years ago, when I've got Noahlink Wireless.

Look at my old gain in bought in 2016 Phonak Bolero V-70P (gains after - ) compared to Infinio Sphere 90 Ultra with P receiver in Unitron power domes (DSL 5.0 Pediatric). My hearing loss was the same:

Bolero, IIRC with skeleton earmold with 1-2 mm diameter vent, this is probably why I didn't get any gain until ~400 Hz.

I don't have interface to connect with my Bolero V70-P, so that graph is only from import from *.tfdata file, which I've got by e-mail. I should have added an so that you could see the first-fit lines onnthe Bolero graph.

@OcramSagev said:

Increasing lows may increase noise and not necessarily lead to better speech understanding

I know that. Plus upward spread of masking mentioned by me. But still, not even a little gain? Especially when vent is small thus better noise filtering.

@Hearing_Potential said:

You can get a much better idea of what hearing aids are doing by using the output view --> real ear SPL.

Yes, I always make a printscreen both gains and output in my PC precisely because of the graph. I think it's a matter of habit. And the feedback threshold graph is missing.

Edited

In reply to @orgullodemexico

@orgullodemexico said:

M receiver

3 wire length

M receiver is too weak for you. will be insufficient to reproduce and differentiate speech peaks from noise.

With your loss, I think UP with very small (<1mm) or no vent at all would be the best option for both ears. Phonak UP receivers has very low noise floor, I don't hear it.

Look at the articles about important role of MPO:

https://forum.hearingtracker.com/t/new-to-hearing-aids-and-this-forum-questions-on-phonak-receivers-domes/102824/20

I also measure my environment.

There's also another factor - it's wire length. I have #0 after my measurements for keeping both mics positioned as horizontally as possible. Look at the wire length part of my topic, although all information is useful:

https://forum.hearingtracker.com/t/advices-before-purchasing-titanium-cshells-titanium-earmolds/106793

In reply to @OcramSagev

@OcramSagev said:

you use DSL pediatric version ( I tested it once and I could not stand it for 1 min). And yes, DSL can prescribe higher values at high frequencies (which I prefer) but many find it uncomfortable.

To be honest, when it comes to maximise performance, I was ready to some trade-offs like "getting used to".

I originally planned to try 5.0 Adult (it is the same as Pediatric, but simply 10 less across ALL frequencies) and add 1-2 dBs every 1-2 weeks.

To my amazement my accustoming period was quite short, maximum few days. I think this is because if you look on my first-fit thin red lines - I am still underfit after 6 kHz:

Looking at the course of thick red lines at the UP receiver, I should also have a few dBs more in the range of 2.2-3.5 kHz.

My first trial with DSLv5 Pediatric was with P receiver with Paradise, which offers even less gains:

According to research papers, first-fit can be even 10 dBs underfit in highest frequencies, so perhaps it shouldn't be a challenge 😅😅

@OcramSagev said:

My bad , I remember when we were discussing the DSL pediatric and I assumed you adopt it going forward.

No problem, I know it's hard to memorise every forum user's history 😅

In reply to @Bimodal_user

@Bimodal_user said:

M receiver is too weak for you. MPO will be insufficient to reproduce and differentiate speech peaks from noise.

With your loss, I think UP with very small (<1mm) or no vent at all would be the best option for both ears. Phonak UP receivers has very low noise floor, I don't hear it.

I've been seen by two different audis and no one has suggested I move to a different power level. I am seeing a third at the end of this month and the question will be asked. Also the wire length of my present set up has never been challenged.

In reply to @Hearing_Potential

@Hearing_Potential said:

One thing that I'm noticing is that people are using the gain curves in Phonak Target rather than the output view. You can get a much better idea of what hearing aids are doing by using the output view --> real ear SPL. If you don't have the MPO curve turned on in this view turn it on in settings.

I'm getting ready to shut it down for the night, but I've spent the past two hours looking at hearing aid output and the client view. I haven't found a guide online on how this mode can be used to ensure whether the output of the hearing aid is adequate; or if this is the right tool for that matter. My background is calibration; use a verified known input to get the expected output. I used two laptops, one running Target and connected to the of course, the other running an online signal generator. I am wearing headphones in a quiet room over the hearing aids; so that the hearing aids only hear the frequency I select. I selected various frequencies i.e. 1Khz, 4Khz and 6Khz just to see how the visual display reacted. My old and tired brain is not understanding how to determine if the output of the hearing aids is correct. I know my setup is flawed as I don't know the amplitude of the signal I am hearing in my headphones. Tomorrow is another day.

In reply to @orgullodemexico

@orgullodemexico said:

I've been seen by two different audis and no one has suggested I move to a different power level.

It's because it seems that they're influenced by Target software suggestion.

It's like in my case, Target suggested M receiver, which was bullshit, and I can confirm that, because I tested also P and UP receivers.

Infinio Sphere with M receiver was worse than older Paradise with P receiver.

Fitting range is a general hint, rule of thumb, nothing engraved in stone.

Please read that articles about . It's also about receiver saturation issue, M receiver is more in risk.

@orgullodemexico said:

Also the wire length of my present set up has never been challenged.

It's an issue rarely analysed. Look here:

Source:

https://www.audiologyonline.com/articles/verifit-directional-mic-test-evaluating-153713

Related discussions