[ThanatosVI]

Hey guys,

I recently started to dab into heavy EQ with convolution filters.
Right now I have fun with EQ'ing one headphone as close as possible to the FR of another with the help of AutoEQ.

I then want to A/B my eq'ed headphone to the original. However I have no Clue how to reliably measure/set the volume.

Are there (affordable) Tools that can help me?
A SPL Meter on a Smartphone yields very unreliable results due to Different seal of different headphones.

 

[dunring]

I use foam to keep the drivers a set distance apart regardless of the pad thickness. Then use a phone app with the mic end the same spot between the drivers. A lapel mic is small enough to fit between them, but sometimes needs a Y splitter to work on a phone.

 

[VNandor]

I'm interested in what others have to say about this topic is well.

As I understand it, in your case the point of the volume would be to make a more fair assessment of how close the EQ'd headphone sound compared to the original. Maybe you even heard about how loudness can affect perceived sound quality.

My two cents is that in general, volume matching the digital files are the most straightforward (for example for fair comparison between lossy/lossless files). It's followed closely by volume matching electrical signals (for comparing DACs/amps). I think it really is possible to get a +/-0.1dB volume match in a lot of these cases, even for hobbyists.
But accurately (or rather, precisely?) measuring sound levels, especially for headphones or IEMs are very hard in comparison...

Something else to consider, a couple of years ago, I wanted to EQ a headphone for a hopefully better sound. I made the EQ profile, tweaked it for some time, enjoyed listening to it but the bug eventually bited me. I wanted to compare my EQ profile to the non EQ'd sound. My EQ obviously made everything quieter so I had to volume match. The question is clear: what property should be measured and then matched so both the EQd and non EQd file would be "volume matched"? I started off by matching the average/RMS levels which made them sound quite similar in terms of loudness but still not quite the same. After some digging, I decided to match the "LUFS levels" instead, which is short for Loudness Units Relative to Full Scale. In short, measuring this unit would take (at least some of) the human perception into account as well, it's not just average levels.

But I still ran into some problems, I measured and matched a ~5min song's integrated LUFS both with EQ off and on but they still didn't sound quite the same in terms of loudness. So I measured the specific part that sounded different and indeed just because the integrated LUFS were the same for the full songs, the readings were different for the specific shorter part I found to be different. Looking back at it, it's obvious that this would happen, but now the implication is that I would have to do my testing by using short samples of songs instead of full songs or even albums. I'm usually fine with that when I'm looking for differences but this time I tried to make a subjective call of which EQ profile I would like more, and using 10-30 second samples for something like that just doesn't work for me.

So in short, I think what's actually meant by volume matching can be ambiguous in some cases and I don't think it's particularly straightforward in your case either. Plus, you also have to deal with measuring headphones as well. Still, might be better than just matching by ears, I really don't know.

 

[bigshot]

I'd say it is probably pretty near impossible to do ABX testing of different sets of headphones because of seal differences as you say. Also, different cans feel different on your head. It's impossible to be completely blind if you can feel it. Switching between one set of cans and another by taking one off and putting another on would take a few seconds, which is long enough to exceed auditory memory for similar sounds.

If you want to compare an EQ curve to uncorrected, your best bet would be to run something like pink noise through and adjust the gain on the EQ to match uncorrected as best you can, either using an SPL meter attached to the ear cup somehow, or to just balance by ear. You aren't discerning if there is a difference. You know the samples will be significantly different and you're looking for a general preference, so it isn't quite as critical.

That said, I don't find switching back and forth to be helpful for EQing. Switching completely disorients me so I don't know which direction I should go. I usually apply a correction curve and listen to a variety of different recordings with it. I make notes of what frequency ranges seem too prominent and then dial them back a couple of dB and listen again. It ends up being like tacking a sailboat back and forth to chart a course halfway between each correction... until I hit the sweet spot and all the recordings sound good and balanced. Bass is the hardest to balance by ear because it varies so radically from recording to recording.

 

[ThanatosVI]

As I understand it, in your case the point of the volume would be to make a more fair assessment of how close the EQ'd headphone sound compared to the original. Maybe you even heard about how loudness can affect perceived sound quality.

Exactly.

My two cents is that in general, volume matching the digital files are the most straightforward (for example for fair comparison between lossy/lossless files). It's followed closely by volume matching electrical signals (for comparing DACs/amps). I think it really is possible to get a +/-0.1dB volume match in a lot of these cases, even for hobbyists.
But accurately (or rather, precisely?) measuring sound levels, especially for headphones or IEMs are very hard in comparison...

Since the sensitivity of both headphones is different, volume matching the Digital Signal doesn't work for me.
However at some point I might have follow up questions on how to do this.(for DAC/Amp comparisons)

Something else to consider, a couple of years ago, I wanted to EQ a headphone for a hopefully better sound. I made the EQ profile, tweaked it for some time, enjoyed listening to it but the bug eventually bited me. I wanted to compare my EQ profile to the non EQ'd sound. My EQ obviously made everything quieter so I had to volume match. The question is clear: what property should be measured and then matched so both the EQd and non EQd file would be "volume matched"? I started off by matching the average/RMS levels which made them sound quite similar in terms of loudness but still not quite the same. After some digging, I decided to match the "LUFS levels" instead, which is short for Loudness Units Relative to Full Scale. In short, measuring this unit would take (at least some of) the human perception into account as well, it's not just average levels.

This is something I thought about as well.
However since my Scenario is to eq 2 headphones as close as possible to the same FR both techniques should yield very similar results.

If I wanted to volume match the un eq'ed headphones I'm back at the same issue tho.
Thanks for sharing.

But I still ran into some problems, I measured and matched a ~5min song's integrated LUFS both with EQ off and on but they still didn't sound quite the same in terms of loudness. So I measured the specific part that sounded different and indeed just because the integrated LUFS were the same for the full songs, the readings were different for the specific shorter part I found to be different. Looking back at it, it's obvious that this would happen, but now the implication is that I would have to do my testing by using short samples of songs instead of full songs or even albums. I'm usually fine with that when I'm looking for differences but this time I tried to make a subjective call of which EQ profile I would like more, and using 10-30 second samples for something like that just doesn't work for me.

I usually use shorter Segments to compare (~30s), so this should be easier as in your case, but still so much to keep in mind for a seemingly simple task.

So in short, I think what's actually meant by volume matching can be ambiguous in some cases and I don't think it's particularly straightforward in your case either. Plus, you also have to deal with measuring headphones as well. Still, might be better than just matching by ears, I really don't know.

My biggest issue is the actual measurement.
How to do it at least somewhat reliably.

I thought about a used miniDSP EARS maybe.

 

[castleofargh]

Hey guys,

I recently started to dab into heavy EQ with convolution filters.
Right now I have fun with EQ'ing one headphone as close as possible to the FR of another with the help of AutoEQ.

I then want to A/B my eq'ed headphone to the original. However I have no Clue how to reliably measure/set the volume.

Are there (affordable) Tools that can help me?
A SPL Meter on a Smartphone yields very unreliable results due to Different seal of different headphones.

In this particular scenario, you're supposed to have fairly similar FR for both headphones, so attempting volume matching does make a lot of sense. Now I see 2 big rabbit holes you can decide to jump into:
1/ if you're going to work out how to measure the signal, chances are that you'd get a better FR match if you just measured both headphones you own with that setup instead of relying on other pairs measured by someone online.
It means more work for you obviously. But also it means potentially a much closer match between the headphones. Here is the giant rabbit hole; If you want to take it one step further and match not just the frequency response but also the time domain from impulses, then I'd suggest to go all the way and get yourself(or make) binaural microphones, the little capsules you stick in your ears. Then you'll get to measure both headphones the way they sound on your head instead of on some random dummy head. You can probably get some advice on those mics in the Impulcifer thread in here that like autoEQ, was made by Jaakko. https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/
That's a deep, hair pulling sort rabbit hole, but also one that can be immensely gratifying once you give up on copying headphones and start copying speakers with impulcifer(you can get an actual soundstage with your headphone!!!! not the 20cm crap with no reverb we call good soudstage on normal headphone listening, real stuff at a distance!!!!! Yum yum! ).

2/ The potential for matching can becomes really good up to a point. No matter what you end up doing, no matter how serious you are and how much money and efforts you invest in this, be it basic matching with a cellphone using one test tone and autoEQ or full on convolution measured at your own ears, the result is unlikely to be audibly exactly identical. With the absolute best approach, there is still some uncertainty with placement, noises, the need for the binaural mics to absolutely not move at all between measurements, and of course the inherent ability of the headphone to be objectively transparent. All headphones have their own distortions profiles and own damping, no matter how good the measurement, the resulting signal is cycled through an imperfect, sometime heavily EQed, transducer. The direct take away is that you could come close, very close. But expecting identical sound is unrealistic. Also, the better the headphone used to imitate another, the closer you can hope to get. HD800 and Stax headphones have been suggested often for stuff like that over the years but newer objectively good headphones with an already pretty smooth FR probably work great too. Which in itself sort of defeats the purpose as what we all want is to use cheap comfy headphones and make them sound like the great and expensive ones. Not the other way around.

Actual answer to your question:
If you want to stick with a cellphone, I'd say play like a 1kHz test tone into your headphones(or another frequency if you see chaos in the FR of one headphone at 1kHz), and use anything that gives you better than 1dB readings. Ideally we need less than 0.1 so that we won't notice and mistake loudness difference for something else but realistically you will get other causes for perceived difference anyway. Testing headphones will be biased and swapping will take too long! It's not going to be the absolute test that means everything even with a solid <0.1dB difference. Just do your best.
Then find some cardboard boxes and make something roughly the width of your head so it can be used to hold the headphones kind of like you would wear it(in term of clamping force at least) while nicely sealing the pads against the cardboard surfaces if at all possible. It's very important if you end up measuring low frequencies, otherwise the impact should be more moderate, just avoid measuring low or high frequencies if there is a giant gap somewhere and the positioning is uncertain(tiny changes could have massive impact on high freqs, but you can just fool around and see for yourself what is stable and what isn't).
try to insert the mic area of the phone near the center of the area where the headphone driver will be. And then while playing the test tone, just adjust the playback level on your audio player.
I don't know for others, but in Foobar2000 you go to "File"->"Preferences"->"Advanced", scroll until you find "Playback" and the specific line "Volume step (dB)" that you can click on to set 0.1. Then you get out off all that and while playing your test tone, you move your mouse over the volume slider and use the mouse wheel to adjust the output in 0.1dB increments until it matches what you had measured on the first louder headphone on your phone app.
Then it's just a matter of writing down those volume levels from foobar and getting back to them when you switch headphones. It will add some more delay on top of changing headphones, but you're already in a not so reliable situation anyway. I think the volume level values is shown on the bottom right of Foobar2000 by default, if not just right click on the bottom line and check "show volume".

The more accuracy you will desire, the crazier you will get with all the extra variables you can't completely control. Welcome to hell, AKA trying to test something properly that involves headphones.

 

[ThanatosVI]

In this particular scenario, you're supposed to have fairly similar FR for both headphones, so attempting volume matching does make a lot of sense. Now I see 2 big rabbit holes you can decide to jump into:
1/ if you're going to work out how to measure the signal, chances are that you'd get a better FR match if you just measured both headphones you own with that setup instead of relying on other pairs measured by someone online.
It means more work for you obviously. But also it means potentially a much closer match between the headphones. Here is the giant rabbit hole; If you want to take it one step further and match not just the frequency response but also the time domain from impulses, then I'd suggest to go all the way and get yourself(or make) binaural microphones, the little capsules you stick in your ears. Then you'll get to measure both headphones the way they sound on your head instead of on some random dummy head. You can probably get some advice on those mics in the Impulcifer thread in here that like autoEQ, was made by Jaakko. https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/
That's a deep, hair pulling sort rabbit hole, but also one that can be immensely gratifying once you give up on copying headphones and start copying speakers with impulcifer(you can get an actual soundstage with your headphone!!!! not the 20cm crap with no reverb we call good soudstage on normal headphone listening, real stuff at a distance!!!!! Yum yum! ).

2/ The potential for matching can becomes really good up to a point. No matter what you end up doing, no matter how serious you are and how much money and efforts you invest in this, be it basic matching with a cellphone using one test tone and autoEQ or full on convolution measured at your own ears, the result is unlikely to be audibly exactly identical. With the absolute best approach, there is still some uncertainty with placement, noises, the need for the binaural mics to absolutely not move at all between measurements, and of course the inherent ability of the headphone to be objectively transparent. All headphones have their own distortions profiles and own damping, no matter how good the measurement, the resulting signal is cycled through an imperfect, sometime heavily EQed, transducer. The direct take away is that you could come close, very close. But expecting identical sound is unrealistic. Also, the better the headphone used to imitate another, the closer you can hope to get. HD800 and Stax headphones have been suggested often for stuff like that over the years but newer objectively good headphones with an already pretty smooth FR probably work great too. Which in itself sort of defeats the purpose as what we all want is to use cheap comfy headphones and make them sound like the great and expensive ones. Not the other way around.

Actual answer to your question:
If you want to stick with a cellphone, I'd say play like a 1kHz test tone into your headphones(or another frequency if you see chaos in the FR of one headphone at 1kHz), and use anything that gives you better than 1dB readings. Ideally we need less than 0.1 so that we won't notice and mistake loudness difference for something else but realistically you will get other causes for perceived difference anyway. Testing headphones will be biased and swapping will take too long! It's not going to be the absolute test that means everything even with a solid <0.1dB difference. Just do your best.
Then find some cardboard boxes and make something roughly the width of your head so it can be used to hold the headphones kind of like you would wear it(in term of clamping force at least) while nicely sealing the pads against the cardboard surfaces if at all possible. It's very important if you end up measuring low frequencies, otherwise the impact should be more moderate, just avoid measuring low or high frequencies if there is a giant gap somewhere and the positioning is uncertain(tiny changes could have massive impact on high freqs, but you can just fool around and see for yourself what is stable and what isn't).
try to insert the mic area of the phone near the center of the area where the headphone driver will be. And then while playing the test tone, just adjust the playback level on your audio player.
I don't know for others, but in Foobar2000 you go to "File"->"Preferences"->"Advanced", scroll until you find "Playback" and the specific line "Volume step (dB)" that you can click on to set 0.1. Then you get out off all that and while playing your test tone, you move your mouse over the volume slider and use the mouse wheel to adjust the output in 0.1dB increments until it matches what you had measured on the first louder headphone on your phone app.
Then it's just a matter of writing down those volume levels from foobar and getting back to them when you switch headphones. It will add some more delay on top of changing headphones, but you're already in a not so reliable situation anyway. I think the volume level values is shown on the bottom right of Foobar2000 by default, if not just right click on the bottom line and check "show volume".

The more accuracy you will desire, the crazier you will get with all the extra variables you can't completely control. Welcome to hell, AKA trying to test something properly that involves headphones.

Thanks a lot for the elaborate input.

Regarding 1)
If a MiniDSP EARS works for that I might do it.
However I read that it does not represent and measure how we hear music, so I'm not sure if comparing different headphones with that device is doable.

Real measurement rigs start in the 5 Digits, and I will not spend that much.

I will look up the binaural recording buds. And Jaakkos project. Very interesting stuff. Thx for mentioning.

2. I'm fully aware that they will never be 100% identical. However I want to see how close I can get with convolution filters.
Especially in the case of Meze Elite and Empyrean, since they have the same earcup shape, material etc. Only difference being the driver.
This is mainly for fun and education, but if I get close enough to justify selling a headphone it's certainly a win.

 

[bigshot]

There was a guy posting here a year or so ago who was measuring various headphones to come up with standardized EQ reference between all of them. He had a dummy head setup and was doing the full shebang. It seems to me, if you are going to try to EQ one set of cans to sound as close as possible to another, you need that kind of a measurement setup. A jury rigged thing wouldn't really prove much because the difference between the two could be due to the scotch tape and foam you use to make whatever it is you use to hold the measurement mic.

 

[ThanatosVI]

There was a guy posting here a year or so ago who was measuring various headphones to come up with standardized EQ reference between all of them. He had a dummy head setup and was doing the full shebang. It seems to me, if you are going to try to EQ one set of cans to sound as close as possible to another, you need that kind of a measurement setup. A jury rigged thing wouldn't really prove much because the difference between the two could be due to the scotch tape and foam you use to make whatever it is you use to hold the measurement mic.

Do you remember a Name, or anything I could feed the search engine with to find that thread?

 

[bigshot]

I don’t remember the name. I’ll search a little tomorrow and see if I can find it.

 

[castleofargh]

Thanks a lot for the elaborate input.

Regarding 1)
If a MiniDSP EARS works for that I might do it.
However I read that it does not represent and measure how we hear music, so I'm not sure if comparing different headphones with that device is doable.

Real measurement rigs start in the 5 Digits, and I will not spend that much.

I will look up the binaural recording buds. And Jaakkos project. Very interesting stuff. Thx for mentioning.

2. I'm fully aware that they will never be 100% identical. However I want to see how close I can get with convolution filters.
Especially in the case of Meze Elite and Empyrean, since they have the same earcup shape, material etc. Only difference being the driver.
This is mainly for fun and education, but if I get close enough to justify selling a headphone it's certainly a win.

I have the mini DSP EARS(got it very early when it came out) and I almost never use it. It's convenient but in term of ear simulation it's crap. There isn't even some old fake 711 coupler in it. The outer ear(I don't know if they changed since but the one I got) is really small in term of inner shapes. Most IEMs with some shell to match the shape of the concha can't fit how they were designed to. So there is no way that ear is anywhere near the average human ear and the FR impact of such ear is ... whatever. The "ear canal" is a straight pipe in the silicon, then the metal plate(only the screws are supposed to ensure proper seal between silicon and metal but deform the silicon and are more likely to break the seal instead...) with a hole in it, and the mic fitted poorly with some foam pushed in to sort of seal that part poorly. At least on mine it wasn't done in a consistent way for the left and right sides.
Some DIY done with any amount of care will easily give better results IMO, even without anything shaped like an ear.

With that said, so long as you're only interested in measuring the FR difference between 2 headphones for your own listening, it's going to be about as good as a real expensive dummy head plus almost as expensive couplers and mics. Because the "ear" used on any device won't be your own so that won't affect the FR the right way for you anyway. And when it comes to FR variation between 2 takes, IMO a mic is a mic is a mic. You start to need real good stuff for distortions or noise measurements(along with a very quiet room!!!). For FR, you just try until the headphone is loud enough without distorting like mad and it's going to give you something you can use to find a direct difference. No point in investing any money for that. The cellphone's mic is already going to do the job(if there is no anti noise or voice clarity dynamic boosting kind of DSP stuck in the middle).
I personally take out the E.A.R.S when I'm feeling very lazy and I just want to confirm something simple real fast. Otherwise I don't.

Here is what I had to say in 2020

Ps: Among my cheap little toys, I have the miniDSP E.A.R.S mentioned by @KeithPhantom. To check variations between headphones(or placement of the same headphone), it does the job. But as some sort of objective standard, the "calibrated" compensations are

It's pretty convenient for headphone comparisons, and pretty bad for IEMs.

Apparently the best I can come up with when discussing the E.A.R.S is say that it's convenient(bold part above^_^ )


IMO, do what you can with what you have, and if you're tempted by Impulcifer, use your money to get a binaural mic and depending on the type of capsule, some cheap ADC with the right plugs. That can do just about anything, including measuring headphone while you're wearing them. Don't hesitate to ask about mic suggestions in the Thread I linked before. There is only a handful of guys in that thread(because audiophiles claim to want the very best sound, but only if it doesn't require actual personal effort to get it). That handful of dudes has 3 or 4 who tried several capsules to make their own mic setup at a various prices. I can't help for suggestions on that as I own a Realiser A16 and have always used the binaural mics provided with it.

Maybe It's my bias because I see Germany, but you're obviously curious and ready to test stuff yourself, so I'm pretty confident that you will want to try speaker simulation, go mad figuring things out, then finally be very very glad that you did.

 

[ThanatosVI]

I have the mini DSP EARS(got it very early when it came out) and I almost never use it. It's convenient but in term of ear simulation it's crap. There isn't even some old fake 711 coupler in it. The outer ear(I don't know if they changed since but the one I got) is really small in term of inner shapes. Most IEMs with some shell to match the shape of the concha can't fit how they were designed to. So there is no way that ear is anywhere near the average human ear and the FR impact of such ear is ... whatever. The "ear canal" is a straight pipe in the silicon, then the metal plate(only the screws are supposed to ensure proper seal between silicon and metal but deform the silicon and are more likely to break the seal instead...) with a hole in it, and the mic fitted poorly with some foam pushed in to sort of seal that part poorly. At least on mine it wasn't done in a consistent way for the left and right sides.
Some DIY done with any amount of care will easily give better results IMO, even without anything shaped like an ear.

With that said, so long as you're only interested in measuring the FR difference between 2 headphones for your own listening, it's going to be about as good as a real expensive dummy head plus almost as expensive couplers and mics. Because the "ear" used on any device won't be your own so that won't affect the FR the right way for you anyway. And when it comes to FR variation between 2 takes, IMO a mic is a mic is a mic. You start to need real good stuff for distortions or noise measurements(along with a very quiet room!!!). For FR, you just try until the headphone is loud enough without distorting like mad and it's going to give you something you can use to find a direct difference. No point in investing any money for that. The cellphone's mic is already going to do the job(if there is no anti noise or voice clarity dynamic boosting kind of DSP stuck in the middle).
I personally take out the E.A.R.S when I'm feeling very lazy and I just want to confirm something simple real fast. Otherwise I don't.

Here is what I had to say in 2020

Apparently the best I can come up with when discussing the E.A.R.S is say that it's convenient(bold part above^_^ )


IMO, do what you can with what you have, and if you're tempted by Impulcifer, use your money to get a binaural mic and depending on the type of capsule, some cheap ADC with the right plugs. That can do just about anything, including measuring headphone while you're wearing them. Don't hesitate to ask about mic suggestions in the Thread I linked before. There is only a handful of guys in that thread(because audiophiles claim to want the very best sound, but only if it doesn't require actual personal effort to get it). That handful of dudes has 3 or 4 who tried several capsules to make their own mic setup at a various prices. I can't help for suggestions on that as I own a Realiser A16 and have always used the binaural mics provided with it.

Maybe It's my bias because I see Germany, but you're obviously curious and ready to test stuff yourself, so I'm pretty confident that you will want to try speaker simulation, go mad figuring things out, then finally be very very glad that you did.

I read about the MiniDSP EARS that the measurements are not really compareable.
E.g. if I measure two headphones and they Show a 5db difference at 5k, this is not a 5db difference I can hear and adjust with EQ.

I probably would have to calculate a compensation curve for each headphone before I can translate the difference in the measurement to the difference I need in EQ.

What does your experience say?

Regarding in ear mics I already did ask for recommendations in that thread

Now that I read it, I'd also like to experiment with the speaker virtualization. However from my current understanding, I'd need a speakersetup first to take the measurements, how it sounds with my ears.
I do have speakers, but absolutely no room treatment, therefore it does sound like utter garbage right now...

 

[castleofargh]

well it really depends on what causes the variation on the E.A.R.S. For example, the "ear canal" can have an absurd resonance(compared to actual ears), but it will be added to both headphones so when you calculate the EQ variation, that particular same resonance on both sides will cancel out.
For stuff like the impact of the outer ear on the waves coming from the driver, things like size of the driver, distance from it or just overall placement can lead to impacts on the FR that aren't comparable on different headphones. In the example you gave, if they both are very similar then that also won't be an issue. But for different headphones it will be. It's the same issue with your own ears and their particular shapes of course, but at least they're your ears so they correlate with how you hear things.


Yes, the result of speaker simulation is at best, only as good as your speaker+room+measurement+headphone+ a little subjective something in your brain and how it handles movements that also move the "speakers". I said not ironically on some thread that I personally happened to enjoy a virtualization of the sound from my android tablet through my headphones more than I enjoy default stereo playback on that same headphone. I cannot know what your situation is and what your priorities are for sound. A majority of people will prefer speaker sound, even crap coming from the TV over headphone playback. But other people do genuinely prefer the cleaner and more detailed feeling they get from headphone sound, even if that comes at the cost of ludicrous spatial interpretation and IMO, a more fatiguing sound. Only you know where you stand on this. If bothering others was no concern at all, would you still spend most of your time at home using headphones?
PS you can always consider bribing a friend or family with a better speaker setup so you can spend half an hour doing plenty of measurements with various mic insertions and playback and recording levels(If you stick to stereo it's quite fast). Some people on the Realiser A16 thread even rented professional studios to get super fancy multichannel measurements. There is no limit to how crazy we can go with that. I've stuck with a simple pair of near field monitors(placed at head level) in a pretty bad room. I do most of my listening in front of the computer so I use that and seeing the actual speakers actually helps anchors the music to them(kind of fun getting fooled by the brain that way).

 

[ThanatosVI]

well it really depends on what causes the variation on the E.A.R.S. For example, the "ear canal" can have an absurd resonance(compared to actual ears), but it will be added to both headphones so when you calculate the EQ variation, that particular same resonance on both sides will cancel out.
For stuff like the impact of the outer ear on the waves coming from the driver, things like size of the driver, distance from it or just overall placement can lead to impacts on the FR that aren't comparable on different headphones. In the example you gave, if they both are very similar then that also won't be an issue. But for different headphones it will be. It's the same issue with your own ears and their particular shapes of course, but at least they're your ears so they correlate with how you hear things.


Yes, the result of speaker simulation is at best, only as good as your speaker+room+measurement+headphone+ a little subjective something in your brain and how it handles movements that also move the "speakers". I said not ironically on some thread that I personally happened to enjoy a virtualization of the sound from my android tablet through my headphones more than I enjoy default stereo playback on that same headphone. I cannot know what your situation is and what your priorities are for sound. A majority of people will prefer speaker sound, even crap coming from the TV over headphone playback. But other people do genuinely prefer the cleaner and more detailed feeling they get from headphone sound, even if that comes at the cost of ludicrous spatial interpretation and IMO, a more fatiguing sound. Only you know where you stand on this. If bothering others was no concern at all, would you still spend most of your time at home using headphones?
PS you can always consider bribing a friend or family with a better speaker setup so you can spend half an hour doing plenty of measurements with various mic insertions and playback and recording levels(If you stick to stereo it's quite fast). Some people on the Realiser A16 thread even rented professional studios to get super fancy multichannel measurements. There is no limit to how crazy we can go with that. I've stuck with a simple pair of near field monitors(placed at head level) in a pretty bad room. I do most of my listening in front of the computer so I use that and seeing the actual speakers actually helps anchors the music to them(kind of fun getting fooled by the brain that way).

Ok from what I gathered, the EARS is not the perfect device for my endeavour.
However binaural mics could do the trick.
Those measurements might only be valid for me, but that's what they're supposed to anyway.

Regarding the speaker vs headphone topic, I always liked headphone presentation and never felt like it's inherenly inferior. In my current Situation the headphone sound I have is vastly superior to the speakersound in my untreated room.
However I instantly had the idea to call dealers to ask if I may take some measurements in their treated showrooms.
Like a setup with the new T+A Solitaire S530

 

[bigshot]

A full size speaker system in a sympathetic acoustic is a whole order of magnitude more amazing sounding than a near field computer setup. It's all incremental... as you add more acoustically balanced space and more sound sources (ie. multichannel), the sound gets better and better. The problem is that it also becomes more complex to balance properly.

The big difference between headphones and speakers is the effect of space on the directionality and timing of the sound. It isn't the aspects that are normally measured (response, distortion, etc.) Headphones are more accurate with those things, but headphones can't provide the kinesthetic effects and immersive sound fields that multichannel speaker systems do. Those are the things that add realistic presence to recorded music. All the things that people imagine with headphones actually exist with multichannel speaker systems (soundstage, depth, dimensionality).