Optimal Frequencies for Speech Recognition

Can’t you simply google up to find out what the speech banana on the audiogram looks like and adjust your amplification to it accordingly?

2 Likes

Thanks. Yes, I’ve done some googling already, so am generally aware of the frequencies involved. That chart you posted is also helpful.

I just thought I’d ask here to ‘pick the brains’ of people who perhaps have already looked into this, who might have some special insight, or specific recommendations, as a result.

As you can see, the speech spectrum is fairly wide, ranging all the way from around 250 Hz up to 4-5 KHz. So amplification adjustment to improve speech clarity is not an optimal way of doing it, because you’d then pretty much have to amplify pretty much the whole spectrum. There’s no one narrow sweet spot just for speech only.

If this could be done like you think, then speech clarity is not such a tough problem after all.

The name of the game is not really to find the sweet frequencies to amplify (because there’s really none, it’s all spread out). The name of the game for speech clarity is to improve the signal to noise ratio (SNR) between the speech signal and the noise. That’s why HA mfgs have to tackle the issues in many different ways, from static noise reduction to directionality for noise blocking and better speech focus, to finding other ways to clean up the speech using dynamic noise modeling to apply noise cancellation principal to speech, etc.

So instead of focusing on trying to find a sweet spot on the frequency spectrum for speech, if you’re a DIY with self programming, try to learn all the SNR improvement techniques your hearing aid model has available, and tweak your programming around those techniques.

Thanks for the insightful reply. Based on what I experienced with the ‘maneuvers’ I spoke of, and with many of the consonants up in the 2k to 4k range, I was guessing that having that narrower range would give me that slight ‘bump’ that seemed to improve intelligibility. So I will play around with those.

But your points about the SNR are well-taken, so I’ll definitely explore whatever means Widex provides to enhance things in that respect. And my initial plan was to create a separate ‘mode’ for movie watching in my main theater, and leave everything else the way my Audiologist has it set, since that environment is quite a bit different from ‘normal’ life!

Thanks again.

I think the short answer to your question is that holding your hands behind your ears does two things. First it reflects more sound into your ears, or in other words increases the amplitude. Second it shields the ear from reflected sound or noise. In the absence of the noise from the back, your ear can turn up the gain more and hear better.

That brings up the question of whether or not your rear speakers in your sound system are helping or hurting. At least in theory they should be helping if the arrival time of the sound is well tuned. However these sources of sound may be competing with reflected sound from the back walls and sides of the room. Some is good because it makes the sound more open and life like, but too much can make it hard to hear. Some hearing aids have an anti echo mode that helps to reduce this effect.

Rather than just changing gains, you may want to explore using a different fitting formula or prescription. There are a number of them that should be built into the software used to set up your hearing aids. The problem however with just changing the formula is that when you change it, then you should have real ear measurements done to verify that in your ear the HA is following the set prescription. Have you had real ear measurements done? It may just be an issue of the HA not being set up properly…

An interesting reply, Sierra.

With your post, and the other discussing directionality, it does make me wonder what the impact is of those rear speakers. It raises the possibility that by having the programming accentuate the dialogue coming from the front, I might lose some of the immersive effect coming from the sides and rear (and above, from the Atmos speakers).

The Marantz AVR I have does included what’s called 'Audyssey XT 32" where you plug in their microphone, and it adjusts timing, frequency response, etc., to at least theoretically optimize its performance for one’s particular room, and should address the valid concerns about this you mention in your post.

But as you point out, the basic room acoustics and reflectivity can be an issue beyond this, and many opt for a wide variety of room treatments to better control these factors.

And if by “real ear measurements” you mean having the Audiologist do a kind of audiogram with the hearing aids in my ear, to test out ‘real-world’ performance, rather than just in the isolation box without the hearing aids, then yes, this is something that he did. If I remember right, this is called ‘Sonogram’ in the Compass software.

He is a good guy, and an experienced Audiologist, and I am reasonably confident that he is doing a good job. My desire for DIY tweaking is not from thinking it’s something he can’t or won’t do, but more that there is just no practical way in an office setting to play around with a variety of settings, nor assess their real-world impact in my home environment.

Can you elaborate on what you mean by “different fitting formula”?

Thanks again.

I have found that wearing ITE custom hearing aids is my answer, I have had one pair of behind the ear aids that were really good aids but I had more issues with noise and speech recognition than I do with my in the ear aids. The sounds are more natural with the in the ear aids too me also.

I think Sierra is on to it. You have all these speakers producing some sound. You cupping your ears lets your ears focus more on what’s coming out the front which is usually the center speaker. This would be the same for normal hearing and aided or not.

Real Ear Measurements are when a small microphone is put into the ear canal first and then the hearing aid. They are used to measure what the HA receivers actually produce for sound levels, and then the audiologist should adjust the gain to match the sound level target. Here is a video that goes into the full detail of it.

The sound level target is set by the formula used for the fitting. The common standard one is the NAL-NL2 standard, but there are other options, including proprietary ones from your hearing aid manufacturer. They vary in the amount of gain and compression applied vs frequency. Here is an article that talks about them.

The newest software to fit hearing aids can automatically do the REM test. When my Audi at the VA setup up my aids he set up the REM test than the software on the computer did the rest

Yes, here is a description of how the Signia Connexx software does it.

You are making a choice to provide a bad hearing environment so unless you prioritize speech in your sound system you will always be making it difficult. Immersive sound is great, but the movie was designed to use wide dynamic range and audio coming from all around you (direct or reflected). In the speech band are also a lot of effects since many effects are broad spectrum. Unless you substantially bring up the level of the center front (or wherever the bulk of the dialog is) compared to the other channels then I think you are living with a bit of denial. Perhaps one day we’ll have selective audio channels in a movie where the dialog can be accentuated in the mix.

1 Like

Yes, the programming software can incorporate/integrate the REM test in it so that the professional wouldn’t have to use their own REM software. BUT, they still need to have the physical hardware equipment to perform the REM test, even if the software is now integrated. Because of that, a DIY like the OP won’t be able to perform the REM test himself without the physical hardware for REM.

That’s the issue with DIY. Without being able to do the REM, any major tinkering, especially with fitting changes or things that will cause a re-prescription on the hearing aid to the new change, will undo the REM adjustment that was done the last time.

3 Likes

So true, when my Audi did the set up the first time and even the second visit he did the REM test twice each time. His comment was better safe than sorry. He seemed to follow the old saying take the time to do it right the first time and it will mean less time spent in the long run. Even when he set up my music and nature programs he spend long than anyone else had ever doing checks

He’s talking about the fitting rationales available to use for amplification prescription. These fitting rationales uses various compression rates at different frequency points and different input sound levels with your hearing loss at those frequency points as an input parameter to determine the proper level of amplification for you hearing loss at a particular hearing frequency and input level.

There are standard fitting rationales like the NAL-NL1, NAL-NL2, DSL, for adult, and DSL for pediatric prescription. Then HA mfgs may have custom proprietary fitting rationales of their own that they prefer to use because they believe it’s best fitted for their hearing aid design as well.

Although you should probably start out using the HA mfg’s proprietary fitting rationale, but if you are still not happy with it, then you can try some of the standard ones to see if you may like those better or not. Some people who are used to a certain fitting rationale over time may tend to prefer to stick with that rationale going forward if they like it better than something new to them.

For what it is worth, I will go in for fitting this week, and I plan to ask for a specific program to watch TV. My system is a 5.1 with rear speakers, but nothing elaborate. What I will ask for is a fixed fairly narrow front microphone focus that reduces sound from the sides and the rear. That way I think I will hear better without compromising the sound that others without hearing loss will have. I should only hear the left, right, centre, and sub-woofer speakers. My hope is that it will have an effect similar to cupping your ears!

They call it Directional Adaptive for the Television program in Connexx for the Rexton KS7’s.
But there’s not really a whole lot of sound that comes out the back speakers. Just environmental ambience sounds or sounds visually moving towards the back. iirc

z10user2:
The TV should have the option, in its sound settings, to choose between surround sound and ordinary stereo.

Agreed. The tv or the STB. But all that sound from the delivered surround sound is getting crushed down into just those speakers and so it becomes a giant mix of sound. Whereas splitting the particular speaker sounds out as delivered leaves the center with just the dialog.

1 Like

I very much appreciate all the knowledgeable replies here. They have been very helpful, and this is why I posted here. A few comments:

Regarding REM, this is how Widex describes the ‘Sensogram’ function in their Compass GPS fitting software:

The Sensogram is an in-situ threshold measurement and one of the most accurate ways of fitting a hearing aid. The measured in-situ thresholds take into account the effect of the earmold/shell, the size of the individual ear canal and the hearing loss.

I don’t know if this serves the same function as REM, but I do know that this was done during my most recent fitting session.

And I agree, that my theater environment is a particularly challenging one. Add to this the fact that the quality and audbility of the dialogue tracks will vary quite a bit from movie to movie, and I realize the inherent difficulties here.

To provide some perspective, the problem I have in understanding dialogue in most movies is not severe, and overall I am quite pleased with the improvement my current aids and fitting provide. I guess I’m just basically wanting to put a little icing on an already delicious cake.

So I don’t feel that I need to reinvent the wheel with regard to my current fitting, but am just curious if a few minor tweaks could add those few extra percentage points of improvement.

I think my initial plan will be to create a new programming slot for movies, with the same settings as what I currently have, and just modify the directionality parameters, to emphasize the front, and de-emphasize rear, and the sides. I do specifically recall him pointing out those options when we did the re-fitting last week, so I know that should be fairly easy to do.

Depending on how that works out, I might explore some of the other options suggested here.

Thanks again to all who have posted their ideas and comments, which I have found very helpful and instructive.