The new method of sound signature copy of any headphones

[outerspace]

I have posted the topic about difference in sound between headphones equalized to one target. If you equalize headphones precisely to one target or to each over and if they not have very much distortions they will sound very similar or most probably the same. You can listen interview of Sean Olive who made similar experiment with same result: "people couldn't really tell the difference between actual binaural recordings of headphones and their only frequency response simulations". So "soundstage", "spatialness", "dullness", "speed" - you name it - seems like all this things are just characteristics of frequency response. If you made artificial ear recording of open headphones and in-ear headphones equalized to them and then compare them with equalized source audio, when all three most probably will sound the same. You can also just pre-equalize recorded headphones to flat and then compare recordings with the source. It would be very interesting to see such test in good realization available for public.

But each time I tried to equalize my headphones using frequency response published on internet resources I don't get good results. Rting, innerfidelty, oratory1990, etc - nothing worked good enough. Maybe it's due to my perceived FR differs from measured ones on rigs. Maybe due to manufacture tolerance. I don't know. But I found a way to copy any FR with amazing result. I was translated frequency response of my open over-the-ear headphones to in-ear by using Y-splitter that split signal from amp to two 3.5mm output. Each output have only one channel - left or right. I insert in-ear earphone to my left ear only and put on open over-the-ear headphones. Then connect both to Y-splitter. So when I play sine wave by generator I hear it only from left cup of my open HP and right in-ear earphone. Then I center this signal on a whole audible range by modifying right channel only (in-ear earphone only) with 31-band equalizer (I used Equalizer APO). I can easy hear left/right channel imbalance on sine signal with accuracy up to ~2-3 dB (depends of frequency range). So I equalize right in-ear to left over-the-ear with 31 band EQ, each frequency at time. If I can catch some additional audible difference I compensate it by parametric EQ. But there is can be a problem is phase differences between L and R on frequencies below 1 kHz. I found method to fix it. Sine generator that I used (SineGen) have phase regulator too so I can tune phase for accurate FR judgment. After equalization I can translate FR from in-ear to any other headphones in similar manner. In this way I can make individual FR copy of any HP. It's not precise, but very good. Method consider my own perception of sound - my pinna and ear canal resonances, etc. Surprisingly, resulted equalization are differs much from FRs published on any internet resources and gave me amazing result. Hope you can get interesting results too if you try it.

UPDATE: You can translate sound of speakers in treated room (studio for example) to headphones by using free software called impulcifer and in-ear mic. After that you can translate obtained personal perceived frequency response from headphones to any other headphones. So it should give similar result to Harman curves but individualized personally for you.

 

[Martin]

I thought you were Amir until I took a closer look at your avatar

 

[outerspace]

Sorry. I just like this japanese hat and Keanu. If it bothers or mislead someone I'll change it. But is this a good representation of your face then you took a closer look?

 

[Chromatischism]

Interesting method. I hope your two ears are the same. Have you had them tested?

 

[solderdude]

I have posted the topic about difference in sound between headphones equalized to one target. If you equalize headphones precisely to one target or to each over and if they not have very much distortions they will sound very similar or most probably the same. You can listen interview of Sean Olive who made similar experiment with same result: "people couldn't really tell the difference between actual binaural recordings of headphones and their only frequency response simulations". So "soundstage", "spatialness", "dullness", "speed" - you name it - seems like all this things are just characteristics of frequency response. If you made artificial ear recording of open headphones and in-ear headphones equalized to them and then compare them with equalized source audio, when all three most probably will sound the same. You can also just pre-equalize recorded headphones to flat and then compare recordings with the source. It would be very interesting to see such test in good realization available for public.

But each time I tried to equalize my headphones using frequency response published on internet resources I don't get good results. Rting, innerfidelty, oratory1990, etc - nothing worked good enough. Maybe it's due to my perceived FR differs from measured ones on rigs. Maybe due to manufacture tolerance. I don't know. But I found a way to copy any FR with amazing result. I was translated frequency response of my open over-the-ear headphones to in-ear by using Y-splitter that split signal from amp to two 3.5mm output. Each output have only one channel - left or right. I insert in-ear earphone to my left ear only and put on open over-the-ear headphones. Then connect both to Y-splitter. So when I play sine wave by generator I hear it only from left cup of my open HP and right in-ear earphone. Then I center this signal on a whole audible range by modifying right channel only (in-ear earphone only) with 31-band equalizer (I used Equalizer APO). I can easy hear left/right channel imbalance on sine signal with accuracy up to ~2-3 dB (depends of frequency range). So I equalize right in-ear to left over-the-ear with 31 band EQ, each frequency at time. If I can catch some additional audible difference I compensate it by parametric EQ. But there is can be a problem is phase differences between L and R on frequencies below 1 kHz. I found method to fix it. Sine generator that I used (SineGen) have phase regulator too so I can tune phase for accurate FR judgment. After equalization I can translate FR from in-ear to any other headphones in similar manner. In this way I can make individual FR copy of any HP. It's not precise, but very good. Method consider my own perception of sound - my pinna and ear canal resonances, etc. Surprisingly, resulted equalization are differs much from FRs published on any internet resources and gave me amazing result. Hope you can get interesting results too if you try it.

I can see this kind of working.
My reservations:
A: you still don't have a target. You are using the FR from the IE as reference. You would need a real reference in one ear.
B: When someone else were to use your EQ it will not suit him because your EQ depends on FR of the IE in your ear canal with your insertion depth and your HRTF on the over-ear headphone.

The reason why 'exact' EQ does not work on ANY test rig has been discussed. Even when they all comply to certain calibration standards and used target responses. Regardless how one is convinced method A using test gear B is the best there is for reasons C under conditions D.

It will differ from individual to individual and from headphone to headphone and even due to production spread of the same model.
Exact EQ does not exist unless calibrated on the individual's head himself.
Even then a small change of position on the head can result in tonal differences.

What EQ can do is bring tonality of different headphones closer together or closer to a personal preferred tonality.

You can not EQ a PortaPro and (fill in your personal favorite high-end headphone) to the exact same tonality and expect the same sound quality at impressive sound levels.
If that were the case it would be a matter of finding the cheapest most comfortable headphone (closed and/ or open) and EQ it acc. to some measurement.
You would only need 2 cheap headphones and audiophile market would cease to exist.

I can go along with FR being the most important factor but it is not the only sound quality determining factor.

IMO the best way is to find a comfortable headphone that needs very little alterations, look at various measurements, look for common 'errors' and EQ (also by ear) the problem areas to either your preference or a reference.
That reference could be a good nearfield monitor or speakers in a treated or EQ'ed room.
That beats making a headphone similar to another non flat 'reference'.

 

[outerspace]

A: you still don't have a target. You are using the FR from the IE as reference. You would need a real reference in one ear.

In this case target is preferred over-the-ear headphones that activates your individual pinna resonances. You translate this FR to in-ear. And after that you can translate from in-ear to any headphones.

When someone else were to use your EQ it will not suit him because your EQ depends on FR of the IE in your ear canal with your insertion depth and your HRTF on the over-ear headphone.

Yes. This method's goal is to make individualized copy.

You can not EQ a PortaPro and (fill in your personal favorite high-end headphone) to the exact same tonality and expect the same sound quality at impressive sound levels.

I see only two problems with portapro. First, they have high level of distortions in lower frequencies and rolled off bass so if you try to fix it you can get audible distortions. Second is FR consistency of on-ear headphones. But if you change PortaPro for Sennheiser HD58X or even something like HD559 you will definitely can get the desirable result. Why not?

If that were the case it would be a matter of finding the cheapest most comfortable headphone (closed and/ or open) and EQ it acc. to some measurement.
You would only need 2 cheap headphones and audiophile market would cease to exist.

Yes, Sean Olive said it works.

You would only need 2 cheap headphones and audiophile market would cease to exist.

For audio transparency you need only decent cheap power cord. But as I can see audiophile market of expensive power cords is not cease to exist.

I can go along with FR being the most important factor but it is not the only sound quality determining factor.

What other factors can you name? And what supportive researches can you point to? Do you know about Olive's and Welti's researches and listening tests? They said that in majority of headphones FR is the only factor that determines sound quality. Do you disagree with them?

That reference could be a good nearfield monitor or speakers in a treated or EQ'ed room.

You can translate your perceived sound of good speakers in the treated room (studio) to headphones by Impulcifer and in-ear mics. And then translate this personal FR from headphones to any other headphones.

 

[outerspace]

Interesting method. I hope your two ears are the same. Have you had them tested?

They are not the absolutely same for sure, but very similar.

 

[solderdude]

In this case target is preferred over-the-ear headphones that activates your individual pinna resonances. Then you transalte this FR to in-ear. And after that you can translate from in-ear to any headphones.

You still do not have an accurate reference. T
The reference will always be 1 headphone that may or may not be correct.
And why the IE ? It is seal and insertion depth dependent. You can also do this with 2 over ears. It just isn't practical.

I see only two problems with portapro. First, they have high level of distortions in lower frequencies and rolled off bass so if you try to fix it you can get audible distortions. Another problem is FR consistency of on-ear headphones. But if you change PortaPro for Sennheiser HD58X or even something like HD559 you will definitely can get the desirable result. Why not?

It looks like you assume you can EQ out cup resonances, sharp nulls etc.

Yes, Sean Olive said it works.

Then it must be true.

For audio transparency you need only decent cheap power cord. But as I can see audiophile market of expensive power cords is not cease to exist.

When all speakers/headphones and electrical devices all would perform similar and everyone could be convinced there is no difference in powercords then we would not see exhuberant prices but only competive prices. Of course there are always folks willing to spend more on thick snakes and nice colors but the whole 'this sounds better' thing wouldn't fly.

What other factors can you name? And what supportive researches can you point to? Do you know about Olive's and Welti's researches and listening tests? They said that in majority of headphones FR is the only factor that determines sound quality. Do you disagree with them?

I agree that in majority of headphones it is the most important factor that determines sound quality. I would not say the ONLY factor. You would also have to understand what the background is of all the research and what 'average' and 'preference' means.

You can translate your perceived sound of good speakers in the treated room (studio) to headphones by Impulcifer and in-ear mics. And then translate this personal FR from headphones to any other headphones.

Then you have a reference to work from. Bombarding one headphone or IEM to a reference is another thing.
From a reference you can use the L-R method to EQ another headphone to your HRTF.
No in ear mics needed nor desirable.
You can also use Griesingers method which basically is the same but uses another principle and ref method but is also usable for personal usage but listening level dependent so should use the method at the levels you normally listen to.

 

[outerspace]

When all speakers/headphones and electrical devices all would perform similar and everyone could be convinced there is no difference in powercords then we would not see exhuberant prices but only competive prices.

Experience said what in many cases you can't convince people in truth. Flat-earhters, anti-vax, covidiots and even audiophiles are good examples of it.

And why the IE ? It is seal and insertion depth dependent. You can also do this with 2 over ears. It just isn't practical.

Because you can easily put two different in-ears or one full sized HP and one in-ear simultaneously. Putting two full sized headphones on one head can be problematic. If you choose correct ear tips, seal would not be a problem, I think. According to measurements from rtings you get very good FR consistency among different in-ears. But I agree it can be a problem in some cases.

It looks like you assume you can EQ out cup resonances, sharp nulls etc.

You can EQ headphones resonances if they not too big, why not? And is the sharp nulls is frequent audible problem in case of even cheap headphones?..

I agree that in majority of headphones it is the most important factor that determines sound quality. I would not say the ONLY factor. You would also have to understand what the background is of all the research and what 'average' and 'preference' means.

I think for majority of headphones it is the only important factor. I made the test where I roughly post-equalized headphones recordings, getting S/N and distortions worse up to 20 dB and even in this case results are very interesting - recordings and source file became very similar. You can download and listen them. I'm sure in case of precise pre-equalization and good ear simulator/ADC you'll get indistinguishable or near indistinguishable result for most of headphones, even cheap ones. Do you disagree with this? Eventually we can test it, I hope. Harman research already done it. It would be nice to see such test available for public.

 

[solderdude]

I know you can get headphones to sound similar but not the same.

Even when you record the best headphone and a poor one and compare these even EQ'd and listeb to the recordings on mediocre headphone you still cannot determine how the better headphone actually sounds. The only thing you can possibly conclude is that they sound similar in tonal balance.

According to measurements from rtings you get very good FR consistency among different in-ears.

On a standard rig with a certain insertion depth... sure.
In practice ear canals differ in shape and how far one can insert and get a good seal is quite individual dependent.
It's the main reason why some find an IE to sound substantially different. And this is not only from taste or tolerances.
Less so with over-ears. On ears and earbuds can also be a hit and miss but measure consistently.

 

[outerspace]

Even when you record the best headphone and a poor one and compare these even EQ'd and listeb to the recordings on mediocre headphone you still cannot determine how the better headphone actually sounds. The only thing you can possibly conclude is that they sound similar in tonal balance.

If you record pre-equalized cheap headphones and good expensive ones and then compare them with source audio and find out that differences between them are negligible why you can't conclude they are actually sound the same with EQ'ing? Do you mean mediocre headphones will mask the differences? I suppose all kind of recorded distortions just stacks with mediocre headphones own distortions so you should hear the difference between recordings and source audio if it audible. At least my own experience shows that only noise floor can mask distortions. And FR can mask or can help to distinguish distortions. I also found if I can detect 1% of some kind distortions then I can distinguish 10% vs 11% too with same easiness. Maybe I'm not quite right here but topic is not straight forward too. After all, you can simply use good headphones for evaluation. I think Sean Olive and researchers used good ones in their similar tests.

 

[JohnYang1997]

Headphones sound different on head than on HATS.

 

[solderdude]

Sonic sense had a whole bunch of headphones they recorded.
Some of those I owned and they sounded different in reality (played on nearfield monitors) because the compensation used wasn't acc. to reality.
You could detect some of the properties and tone though.
Oluvs attempt was better.

Do you mean mediocre headphones will mask the differences?

exactly, you can't expect a 'normal' headphone to reproduce what some 'better' headphones can. It is not all FR, just mostly.

The phrase you can't make a silk purse from a sow's ear is also true for headphones.
For untrained and unsuspecting listeners that don't know what to look for sure... 2 similar EQ'ed headphones will certainly sound the same.
When I let a family member listen all they can say is whether they like it or not or if its comfortable.
They can hear a good headphones sounds 'better' but not where they excel.

 

[outerspace]

For untrained and unsuspecting listeners that don't know what to look for sure... 2 similar EQ'ed headphones will certainly sound the same.

The point is you can't precisely equalize headphones to same FR by just using graphs from rigs. They will sound different for sure. Even just variables in positioning can make them different. It's not the thing what I'm talking about. I'm talking about good realization of test with headphones recordings such as Olive made.

you can't expect a 'normal' headphone to reproduce what some 'better' headphones can

For me it's just your word against researches of Harman lab. Did you made some test too? What are your conclusions based on? Did you interested to made similar test?

 

[maverickronin]

EQ can make huge improvements in most headphones, but it's not everything. "Basic" FR adjustments can only get you so far.

For example, the soundstage and sense of space in headphones comes mostly from the driver's interaction with your pinna. Drivers that are larger, angled, or farther away tend to produce larger soundstages for this reason. This is still an effect of FR and can be emulated, but it will require personalized measurements of both headphones on your own head and full on convolution. Measurements from a generic simulated pinna won't cut it.

 

[outerspace]

For example, the soundstage and sense of space in headphones comes mostly from the driver's interaction with your pinna. Drivers that are larger, angled, or farther away tend to produce larger soundstages for this reason. This is still an effect of FR and can be emulated, but it will require personalized measurements of both headphones on your own head and full on convolution. Measurements from a generic simulated pinna won't cut it.

I agree. But it's not that I'm talking about.

 

[solderdude]

I'm talking about good realization of test with headphones recordings such as Olive made.

He just used his HATS for the recordings with a different target response most likely. Just like Sonic sense did and Oluv.

I have experimented with recording sound as well but for well known reasons it did not mimic the sound on flat monitors.
There are too many aspects besides FR.

I do like your idea to make headphones sound the same which is simple.
The issue I have is reference... it all hangs with reference when you want to create something meaningful or you want to headphones to sound very similar. If one of them is reference grade then it makes sense.

 

[outerspace]

He just used his HATS for the recordings with a different target response most likely. Just like Sonic sense did and Oluv.

Harman used GRAS. Oluv's rig is not high grade. Sonic Sense used just microphones, not even artificial ear. And for test one need to precisely equalize headphones to flat before recording. Most valuable conclusions we can make from recordings made on good measurement rig with ear simulator such as GRAS. But even Oluv's rig would be interesting if he would made precise equalization before recording. I contacted him and he said something like "it'll sound the same nothing interesting". I contacted oratory1990 but he didn't answer. Later I found he wrote on forum something like "if you equalize headphones precisely then yes, they will perform identical. But precise equalization on real human is not possible, there is always be some little differences due to different positions, hairs, etc".

 

[solderdude]

Sonic Sense used just microphones, not even artificial ear.

No they used a HATS for the headphones recordings. Oluv's rig has incorrect compensation (I assume he used the compensation that came with it) so he could EQ acc. to his measurements and would sound accurate but not correct in an absolute sense.

It's all about reference ! Even HATS aren't the same reference and differ and differ yet again from reality. Sure, they comply to standards (for different usage) and you can put a headphone on it and get measurements. Doesn't mean they are accurate in an absolute sense.
It is will be better than nothing or cheap in certain aspects.

 

[outerspace]

HATS would be interesting. Compensation curve doesn't matter in case of this test. We take some source audio and record it through the chain DAC>HP>RIG>ADC. In the result we should get a recording with exact the same frequency response as source audio. For this we just measure FR of all chain and then apply minimum phase EQ before recording process to make it flat as source audio. The most important thing here is mic and ADC with highest possible S/N ratio for given headphones SPL. In such test we can compare headphones recordings, precisely equalized to flat, with non-modified source audio. So we can compare all differences besides FR. I think it's can proof am I wrong or right about FR is the only important thing in majority of headphones. If we don't take into account Sean Olive who already proof it in Harman lab.