Custom EQ to address mild hearing loss in left ear

[Dougey_Jones]

Hi Everyone,

I've suffered from a different FR in my left ear for many years at this point and am interested in doing something about it, but I need the assistance of ASR in order to execute it properly because I don't have the methodology hammered out in my mind just yet.

Basically, my left ear has mild hearing loss, probably 3db'ish compared to the right, but the 3db is not uniform, so it's not easily addressed by simply tweaking the balance to the left. I end up having to change the balance on every track, which is annoying. It's also much harder to address for Home Theater, since there are Front Left, Surround Left and Center channels to consider, how would I even EQ a center channel to compensate for this problem?

My thoughts so far are as follows, and I would invite anybody to assist me in fleshing this process out so that I can succeed. It may end up being helpful for others suffering the same type of hearing loss.

Proposed Methodology:

1) Use individual test tones, one at a time and note how many db "off center" each tone is.

2) Accumulate this data in a spreadsheet or notepad.

3) "Combine" these measurements with the Harman curve, which I personally enjoy very much. The standard Dirac Live 3 FR curve that I've used with my MiniDSP Flex is very enjoyable and basically the Harman Curve, so that FR curve +/- the "correction curve" to adjust for my left ear's deficiencies will make up the target FR curve for the left channel, surround left channel, etc.

4) Load these target curves into my MiniDSP for the stereo signal chain of my system.

5) Figure out how to load the same curves for the Left + Surround Left channels for my Denon x3700h, using the MultEQ-X software. I would stick with the MultEQ editor app for iOS, but I don't think the iOS app allows for importing target FR curves. I don't mind buying the MultEQ-X software it it helps me accomplish my goals.

There are some really talented people on this forum, I'm hoping some will take an interest and help me muddle through this process so that I can finally address these issues for myself, it'd be truly remarkable if I didn't have to hold the preamp remote in my hand and adjust the balance for every new song. The biggest challenge I see is gonna be addressing the center channel dialogue seeming slightly off center (to the right), because it's a single channel being produced by one speaker that is physically centered in the room. If I were using two speakers to create a matrixed center the problem would be more straightforward to address.

Any help greatly appreciated.

 

[Timcognito]

Hey Dougey, I went through this with my mom and the easiest way to deal with it was go to Costco and get a free test by a trained expert, plots for both ears and learn about hearing aids. They carry many inexpensive up sophisticated and expensive ones ($2500 in her case). If you decide not get anything you get your report showing what frequencies and differences ear to ear, real data to tinker with.

 

[Dougey_Jones]

Hey Dougey, I went through this with my mom and the easiest way to deal with it was go to Costco and get a free test by a trained expert, plots for both ears and learn about hearing aids. They carry many inexpensive up sophisticated and expensive ones ($2500 in her case). If you decide not get anything you get your report showing what frequencies and differences ear to ear, real data to tinker with.

I actually had a test performed by an audiologist that shows the differences in my hearing FR, I just thought that conducting my own tests using audio equipment rather than IEM's in a listening booth would produce a more satisfying final product than creating EQ curves based on the audiologists results. It would allow me to test at a minimum per octave or even more fine grained depending on how many test tones I use.

I don't have any plans to get a hearing aid at this point, I'm in my early 40's and the hearing loss in my left ear is an annoyance for an audiophile with OCD, not a "real" problem.

 

[olieb]

Proposed Methodology:

1)-4)

This would be exactly how I would do it for headphones. And of course Tim is right that it is probably best to get a audiogram from an expert with a proper instrument first. If this does not do it you can still go the other way.
But there is a problem with EQing speakers this way.
It will probably not work so easily. You already mentioned these problems when considering the center channel.
If you change the FR of the left channel this will change the sound going to your right ear too!
The relation of left and right channels for phantom sources depends on the recording (miking) technique. A panned recording has different levels in the channels, a stereo recording with AB technique (omni mics) will have much less channel level difference. And therefore the resulting recordings can be all over the place. So perhaps you will be stuck with the remote.
But give it a try.

 

[Dougey_Jones]

If I posted my audiogram would the forum be able to assist me in making target FR curves using that data?

EDIT: Thank you Olieb for pointing out my obvious oversight that EQ'ing the left speaker to compensate for my left ear FR would also affect my right ear, thus making this less straightforward than if we were using headphones. I still think it's worth trying, and may be "better than nothing". The same basic principle should also apply to room correction, where EQ'ing any single speaker's FR will affect the overall FR of the whole room. Yet we still use room correction.

 

[AudioJester]

Have a look at the hifi aid on this site

PLUGINS

Digital room correction plugins

www.homeaudiofidelity.com

Maybe helpful for you

 

[Dougey_Jones]

Have a look at the hifi aid on this site

PLUGINS

Digital room correction plugins

www.homeaudiofidelity.com

Maybe helpful for you

Wouldn't this be hard to integrate in a "blanket" way so that it affected all sources? If I integrate the corrections into my AVR using MultEQ-X and my stereo signal chain via Dirac Live 3 or via the MiniDSP application then I can expect it to be applied to all inputs on those devices.

 

[olieb]

The same basic principle should also apply to room correction, where EQ'ing any single speaker's FR will affect the overall FR of the whole room. Yet we still use room correction.

Sure, but that is why it is called room correction, you correct the influence of the room. But having two ears is like having two rooms. Now if you apply room correction it might be a good fit for one room, but the other will get the same "correction".
As I said, it might work to a degree, but it will depend on the recording (technique) and will have its flaws.
To a degree your brain should adapt too, getting used to having less "loudness" (nerve stimuli) from the left ear when the source is in front and correspondingly for other source positions. And all that depends on frequency range of course.

 

[Dougey_Jones]

This is what we're workin' with. What's interesting, is that when I adjust balance, or in the case of the AVR, boost the left side of the room to compensate, I haven't ever adjusted more than 3-5db. From the Audiogram you'd think that I'd need 10+ DB of compensation for it to be perceived as "centered" but that hasn't been the case, so maybe my brain is compensating to a certain degree.

 

[kemmler3D]

I would tend to agree that single-ear corrections for speakers might be a nonstarter because of crosstalk and reflected sound. Consider that at the critical distance, a single-channel correction will be (depending on how you look at it) more than half wrong, and that's only about 2-3 meters in most rooms.

I would experiment with minor changes to correct FR but I think you should approach it as a change you're making to both channels for the combined response of your ears, if that makes sense.

Pretty doable for headphones using the proposed method though.

 

[Timcognito]

After thinking about it, correcting your ear and your room separately might be easier and more effective. But I understand, in my forties I would have put off hearing aids, but some inexpensive one(s) for personal listening, might be worth a try. Food for thought

Pretty doable for headphones using the proposed method though.

This is the same or similar what I am suggesting.

 

[olieb]

Ok, this is not a lot of resolution. 2values/oct and 5dB steps.
I do not know about these audiograms, is this the threshold of hearing?
Or is it like an equal loudness at x phon?
In any case I would claim, that I can do better for the situation of music hearing with headphones at "my" loudness and test tones. A relative resolution of 2dB for left/right balance is readily achievable.
My favourite test tones are (mono) bandwidth filtered (with BU8) pink noise in 3rd octaves from REW.
Then adjust left channel for center perception.

EDIT: I will join the choir and recommend headphones. It will work much better and sound is better too. You might want to check out the binaural virtualization thread.

Threads of Binaural virtualization users.

Across global audio communities—including ASR—there are listeners who use headphones or IEMs to recreate a speaker setup via binaural virtualization. Although you can find bits of information and discussion, it’s scattered and often hard to track down. And even within audio forums, the typical...

www.audiosciencereview.com

EDIT II: It just came to my mind that you can use these test tones over speakers too. You can check if you perceive a mono source in the center between the speakers or not and in which frequency ranges there might be what kind of deviations.

 

[Dougey_Jones]

Acknowledging that doing this for headphones would be easier and more effective, I'd still love it if someone will help me calculate "combined adjusted" target FR's that are exportable to Dirac Live / MultEQ-X that either bring my left ear in line with my right ear's FR, or correct both ears to flat. The trouble doesn't really start until 1khz, I don't think any compensation is necessary for either ear until then.

@Maiky76 came to mind, since he seems to breeze through the EQ recommendations for all of Amir's speaker reviews, but I'll take any help I can get from anyone.

 

[Timcognito]

Acknowledging that doing this for headphones would be easier and more effective, I'd still love it if someone will help me calculate "combined adjusted" target FR's that are exportable to Dirac Live / MultEQ-X that either bring my left ear in line with my right ear's FR, or correct both ears to flat. The trouble doesn't really start until 1khz, I don't think any compensation is necessary for either ear until then.

@Maiky76 came to mind, since he seems to breeze through the EQ recommendations for all of Amir's speaker reviews, but I'll take any help I can get from anyone.

Okay suggesting practice on HP as that would help set up a target as both your ears and then whats missing in the room that needs work. Plus now your HP listening is better too. I know I get more nuance from them.

 

[olieb]

I'd still love it if someone will help me calculate "combined adjusted" target FR's that are exportable to Dirac Live / MultEQ-X that either bring my left ear in line with my right ear's FR, or correct both ears to flat.

If you base that on the audiogram from above (which I would not do) then this is the difference between your ears:

If we ignore the value from 250Hz it looks consistent though "stepped" from the bad resolution. The green curve would be a start to create a filter for the left ear (relative to the right).
In REW for Minidsp (flex) the implementation could look like that with high shelf filter (adjustable Q).

But I will repeat: I strongly doubt that this will work. This will increase the level of i.e. 3kHz in the left channel by 10dB!
I cannot think that this will sound good in any way. Your right ear will crawl up the wall.

Just try the test tones from above first. My guess is that if anything then the compensation will look very different than what the audiogram implies.

 

[Dougey_Jones]

It has to be possible to arrive at an in room EQ (at least for L/R) that makes most frequencies sound centered, otherwise my using the balance control on my pre would do nothing for me.

Maybe the play is running Dirac and Audyssey first with the Harman target curve and then playing a test tone centered on the available PEQ points for both my MiniDSP and my x3700 and manually adjusting the PEQ for each device to establish a balance at each PEQ point from my main listening position.

IDK if the Denon has that ability, or what PEQ intervals are available but it should be possible in the MiniDSP software at least.

 

[kemmler3D]

It has to be possible to arrive at an in room EQ (at least for L/R) that makes most frequencies sound centered, otherwise my using the balance control on my pre would do nothing for me.

Maybe the play is running Dirac and Audyssey first with the Harman target curve and then playing a test tone centered on the available PEQ points for both my MiniDSP and my x3700 and manually adjusting the PEQ for each device to establish a balance at each PEQ point from my main listening position.

IDK if the Denon has that ability, or what PEQ intervals are available but it should be possible in the MiniDSP software at least.

Agree that an empirical approach is probably the easiest way forward here. Since both ears hear both speakers, I can't predict what the subjective outcome of EQ on one speaker will be.

Rather than tones I would suggest playing pink noise (maybe band-limited if that helps) and EQing that. Tones are good for figuring out your hearing/response at specific frequencies, but when you're subjectively assessing tonality, PN is WAY better than tones. In-room, tones are also impacted by standing waves in a way that PN (subjectively) is not.

Especially since we're talking about frequencies that are smaller than your head at some point, judging whether a pure tone is louder in one ear or the other in-room, especially when hearing loss is a known factor, will be verrrrrry slippery. Because it actually will be louder in one ear than the other. Or is it? Move your head and it could change by 6dB no problem. A good way to drive yourself crazy and come up with a random EQ setting that doesn't help.

This is the same reason people are advised to use MMM or multiple measurements in-room if they're doing anything with correcting higher frequencies.

I would also start by getting the L/R balance just right first, since adjusting level after adjusting EQ will tend to throw off the results of the EQ via loudness effect.

I think the goal would be to be able to play pink noise over the L, R and both channels and have it sound subjectively about the same in each case. I am not sure if that's possible... but I am interested in hearing how it goes.

 

[Dougey_Jones]

Agree that an empirical approach is probably the easiest way forward here. Since both ears hear both speakers, I can't predict what the subjective outcome of EQ on one speaker will be.

Rather than tones I would suggest playing pink noise (maybe band-limited if that helps) and EQing that. I would also start by getting the L/R balance just right first, since adjusting level after adjusting EQ will tend to throw off the results of the EQ via loudness effect.

I think the goal would be to be able to play pink noise over the L, R and both channels and have it sound subjectively about the same in each case. I am not sure if that's possible... but I am interested in hearing how it goes.

This should be very possible with my MiniDSP Flex, and will be my first goal to make it happen with my Flex for stereo signal chain.

After researching, the x3700 doesn’t have its own PEQ per channel, so it sounds like my only option is to make it all happen with Audyssey, which should still be possible, but will be a much more involved task, including not only figuring out how to map pink noise tones to my surround L/R for testing but how to adjust the balance for testing. Definitely not possible with the iOS app, but maybe doable with a ton of effort in the MultEQ-X windows software,

Are there any MultEQ-X users that can chime in regarding whether the software allows you overlay PEQ on top of your target FR curve on a per channel basis?

 

[kemmler3D]

figuring out how to map pink noise tones to my surround L/R for testing but how to adjust the balance for testing. Definitely not possible with the iOS app, but maybe doable with a ton of effort in the MultEQ-X windows software,

For developing the EQ and messing with test tones, if you have a laptop I'd highly suggest using that as your source and just using REW and/or EQAPO/Peace or even just Audacity to do realtime tweaking. Alternately you could even set up a DAW (FL Studio has a free demo) to have more control over everything for testing, then translate the resulting EQ to MultEQ-X.

Unfortunately I have not used MultEQ-X so I can't advise on setting that up once you have the intended adjustment locked down.

 

[Dougey_Jones]

I think I have the plan locked down in my head for MiniDSP signal path.

Hopefully a MultEQ-X user will show up and advise before I spend $200.