I’m a musician, I do mixing and mastering in an ATMOS room, I’m an electrical engineer, and I need a hearing aid. This is just a short list of my complaints about the hearing aid manufacturers in general. I’m hoping someone in their origination sees this and feeds it back to their development teams.
-
Web sites:
The “customer” web sides of most HA manufactures is a glossy vague description of how wonderful their hearing aids are and how they will make you hear better than ever. Fine. That’s about the level most people can relate to. However, if you have some background in music, acoustics, recording, etc, you’d really like a bit more details. So of course you can go to the manufacturer’s “pro” web site which has the glossy datasheets that audiologists like in that they show the styles and colors and features of the hear aids offered and some level of information on fitting range. Having spent some time recently on several of these sites, the “documentation” still fails to offer the basic information that I’m interested in as a professional consumer. -
Specs
So what’s missing. Things I want to know that are very difficult to find (sometimes an has some answers, but do you trust them, and even they fail on most basic questions). Asking an audiologist has gotten mixed results with them occasionally trying to get the information from the manufacturer. Since I’m shopping for a new HA right now, I actually called into my two primary contenders support teams and managed to get through to one of their audiologist. I was lucky to have a long conversion with one manufacture’s audiologist who was actually a musician, but many of my questions were not answered with a definitive response. So here’s a list of what I’m looking for:
a) The pro specs list the number of channels (bands) their various offerings support. The premium ones are in the, say, 15-24 channel range. What are the center frequencies of the bands? Having had a hearing aid in the past (“one for musicians that has 15 bands”), I’m aware of what their fitting software shows for the bands. They make sense in that they scale logarithmically and mostly fall on standard octaves or 1/3 octave frequencies (in my case - 125, 250, 350, 500, 630, 800, 1K 1.25K, 1.6k, 2.0k, 3.2k, 4.0k, 6.0k, 8k - which they spec as covering 100-10KHz which is about right given the -3dB points would be about there). This is pretty sensible as there are more or less evenly spread through the available octaves (e.g. 100-200 has 1. 200 - 400 has 2. 400 - 800 has 3, 800-1.6k has 4, 1.6k-3.12k has 3, 3.2k to 6.3k has 3, and 6.3k to 12.5 has 2). Another brand that had 24 bands had 150, 500, 750 1100, 1400 1700 2000 2300 2700 3000 3300 3600 3900 4200 4500 4800 5200 5600 6900 7500 8100 8800 9400. This is more or less linear. so the first two octaves (100-400) have just 1 1. The last octave 6.3k-12.5k has 5. This is a bit silly. there’s not enough resolution at the low end and way to much at the high end where more than a few people can’t hear anyway. If they actually told anyone about this, I could see where I general consumer wouldn’t know about octaves enough to understand why logorithmic makes sense, but they don’t talk about this. I know which of these two I would choose. I have some very GROSS changes in my hearing between say 100 and 500 Hz. I need more control in that range.
b) What is the maximum latency? Under what program setup conditions. Hard to get this detail from anyone except one vendor, in my experience. For a musician that plays an instrument, over 5 ms is getting into the audible range if you can play fast - as in you can’t play fast with latency greater than this. If you have an open dome, comb filtering is going to cause artifacts at f = 1/latency. So at 5 ms, your combing at multiples 200 Hz. At 1ms your combing at multiple of 10,000k. Overall, comb filtering is not overly noticeable - but if you’re trying to tune in a mix at some frequency, best not to have a comb filter from the hearing aid (which might be in addition to any comb filtering your speaker placement is already causing). So lower latency is better - get it up as far as possible above your range of hearing ideally. A workaround, use a to avoid combing. But at least tell me what your HA does.
c) Can you turn off dynamic changes to anything in a particular program? Anything is noise suppression (dynamic wide band compressor), environment adaption (dynamic narrow band compression/filtering), focus adaption (modify rear or side response levels), anything AI. In a mix room, there is NO noise. Anything that AI thinks is noise is part of the recording mix. It needs to be heard at the level it is recorded at. Explosions, 0 to Max dB transitions, surround speakers at ear height and ceiling height, bass response down to as low has the HA covers (and below). I was amused by one implementation of noise suppression. For something like a transient of a coffee cup hitting a table, you hear (1) The coffee cup impact at full volume. (2) 200-300 ms later you hear the compressor kill all the HF response of the HA, (3) at 2-4 seconds later, you hear the compressor open up the HF response of the HA. So every transient becomes three transients. Nice job! I sure if you are on an airplane, it might give you want you want, but in general, it’s not a good thing. What’s particularly bad here is the HA manufactures are all jazzed about their AI adaption to manage hearing speech in noisy environments. That’s great and is useful when you’re dealing with conversations, but let me TURN THIS OFF in a studio/music environment. “But we have a Music Program”, you say, “and musicians love it”. Of my two shortlist HA’s, one’s audiologist said you could turn off 'almost everything" - didn’t get a clear answer of what she meant by “almost”. The other, which is in the ‘early release phase’ of their new platform has limited support in their fitting program at this point and the audiologist said you can’t ‘turn off’ the AI. I’m not sure I totally believe her or know what their music program actually can be configured for, but this trend is disturbing.
d) I can’t see the fitting program software to figure out what it can do. I’m not an audiologist - not allowed to see it. In some cases, I could download it illegally through 3rd party to have a peek. In other cases, I can’t do that even. Or I could buy the HA so that my audiologist could show me the fitting program (and then return them if I don’t like what I see), I understand most people wouldn’t care about this. But given that, why restrict me from seeing it? It’s ot like it’s a competitive secret. It would take the inclusion of a “demo” mode to let me in to see the panels without having the HAs.
e) What’s your input dynamic range - and what is the maximum input dB. This is not to be found in the HA pro specs I looked at. Both manufacture’s audiologists seem to eventually agree with the dynamic ranges I was giving them (one was 108 dB, one was 108? or 118? dB). Neither knew the max or min, but one said that the latest HA hadn’t changed that spec from the prior generation and I could find a spec (from AI) on that one. Why do I care? Well in an ATMOS mix room with an 85 dB-C reference level. One non-LFE channel peaks at 105 dB-C. But the mix is limited by the loudness of a 5.1 mix down so that 85 reference ends up peaking at about 120 dB-C SPL. This is a worst-case situation and more likely something like an explosion so does it matter if there is some transient distortion? I’d rather not have to worry about that. Fortunately, my mix environment is not at 85 dB-C, and I’m looking at maybe 111-114 dB-C. Would it kill you manufaucturers to have a 21 or 24 bit ADC versus whatever you have now?
f) Do you do any testing for directional accuracy. So I know everyone’s hearing is different, but lets say you program your HA for someone with normal hearing and then have them listen to a test signal that is routed through an ATMOS monitoring (like 9.1.4, not binural on HP). Can then tell the positioning of the test signal as accurately as without the HAs in the full 360 degree ear level positioning? Can the tell the location of the test signal on the height speakers as well as without the HAs? Some of the answer to this will have a lot to do with where you put your pickups and how you manage them. But some sort of technical report on this would be very helpful and interesting. It might even convince you to move your pickups.I don’t really care how it looks on my head. I want the hear what is there. While we’re on ATMOS, I could make the same comment about testing directional accuracy with bi-naural audio on headphones (but also see the next comment on HPs in general).
g) Do you do any testing with HA and headphones. I’ll convey my experience with a HA. I use double domes because otherwise I get feedback with the HPs on. I’ve noticed for a long time that headphones are almost not usable.I have one bad ear and one almost normal ear (no HA). It may be my fitting was way off, but the levels on signals that are balanced full left are way low. Experimenting, I held the piece sideways on my pinna (facing the headphone speaker) and the response levels were normal. So it seems being in the shadow of the pinna is blocking the signal. (Music example - intro to Stairway to Heaven has the acoustic guitar balanced full left). Two other test points I noticed around this time: 1) Watching TV with a decent stereo system sitting on a couch near the rear wall, I was picking up the rear wall bounce much more than the direct signal. Moving the BTE piece to face forward fixed the response. When sitting in a restaurant or coffee shop, it’s a lot easier to hear the conversations going on behind me than the ones going on in front of me. I’ll mention that I’ve always left my HA set on a Music program where we turned off everything that could be found to turn off. One vendor has a RIC with a pickup on the receiver in the middle of the pinna and I was interested in that, but I have not tried it and probably there were other reasons that HA fell off my short list, but again, placement of the pick ups is poor.
h) I’d like to understand more about any limiting done on the output chain. I’ve had the sense my HA has pegged a limiter, but it might be my hearing. Perhaps a signal flow diagram with major components would be helpful. I know there’s a lot in the digital domain, but something descriptive would be better than nothing.
-
I’ve not been happy with my fitting process. Last HA, audiologist had no decent speakers, my HA doesn’t work with a HP, and my HA didn’t stream. I had to got get the fit tweaked, then test it in the studio, and then figure out how I wanted it changed, and then repeat with the audiologist trip (a lot). Three potential solutions here: (1) Audiologists, get a decent sound source. (2) use streamed audio during the fitting process - better if don’t have to disconnect HAs while trying out hte streaming. (3) Support remote fitting. It’s pretty common. But one manufacture just released a new platform that doesn’t have it. Couldn’t get an answer about if they had plans to eventually support it. I’m a bit sympathetic in that I’m aware of the problems with per-disclousre of unannounced functions that may or may not show up an have an undetermined development schedule or commitment. However, something this basic perhaps should be available in the first release phase. I can say it may put me off buying this HA even though it is otherwise my first choice. I could deal with it not being there now, if I knew it was in a future release plan. It probably is. But getting anyone to tell me this from your customer support team has been unsuccessful.
-
I haven’t run into this myself, but I’ve heard others complain about having pitch issues with their HA. I’ve read about at least one feedback management program that exploits detuning. I"d be suspicious of anything like this in the 100-1K range where a few hertz is more cent’s off. Make sure it can be turned off if your doing this.
-
I’ve had problems in the past with the “factory” fitting being pretty awful. I’m not sure that going from an that measures the edge of hearing to normal SPL levels tracks well. I guess this comes down to sort of an opinion survey of what customers thing. And everyone’s hearing problem is different so it’s hard to get much of a statistically accurate reading. But my point is that I think folks would probably want to tweak their fitting a lot more than audiologists know if they had some way of experimenting on doing some self tuning. I’m aware of some of the perils of “” programming, but something between what is current best practices and something that relies a bit more on customer’s ability to adjust their fit might be warranted. I’ll call this “Directed and Monitored DIY” for grins, but it could be an interesting area of research.
Anyone else have thoughts?
You’re ignoring this member’s posts.
Hot Springs Village, Arkansas
Canada (Toronto) YYZ
Taiwan
especially now that most can stream bluetooth audio.