Headphone Measurements Using Brüel & Kjær 5128 HATS

[Mad_Economist]

Listen to speakers.

Of course, if we look at the eardrum response on humans with speakers, we'll see some substantial unevenness at even lower frequencies...

But for higher frequencies, particularly in direct sound, this definitely is a quite reasonable argument.

 

[Blumlein 88]

Of course, if we look at the eardrum response on humans with speakers, we'll see some substantial unevenness at even lower frequencies...

But for higher frequencies, particularly in direct sound, this definitely is a quite reasonable argument.

Very true, but more amenable to mostly solving problems below Schroeder frequencies. There is always get a bigger room or listen outside.

 

[Tks]

Listen to speakers.

One problem, that's not a solution for headphones :\

It's not terribly hard to measure 10khz - you can acquire microphones, amplifiers, and ADCs with sufficient bandwidth for the task for pennies these days.

The "problem" is that when wavelengths are small relative to the structures of our ears and the dimensions of the volumes we're dealing with, things simply do vary, and measurements on HATS accurately reflect this. IMO considering this problematic is sort of like pointing to a plot of many different axial HRTFs and saying "Why do these vary so much when you're paying $20k+ for this thing??".

The variation in high-frequency response is a real phenomenon on our actual ears and at our actual eardrums, and you're not likely to get past that without some truly invasive measures.

Oh okay, so no solution, just a literal limit of reality of the matter then, correct?

So the "problem" gets worse with smaller structures, or what's going on here? Would it then be safe to assume IEM's are the worst offenders (meaning introduce the worst examples of variability above 10K), or do they recoup a bit considering they bypass notions of cancellations and such other things that I assume would be more prevailent within the space of a headphone cup?

 

[kn0ppers]

Get some Sony 7510 ro 7520 phones, they have big roomy earcups.

If you can find one! The 7520s and their HiFi counterpart Z1000 have become rarities.

Edit: As I say that, I find some on a well known auction site. Interesting. I might have to pull the trigger on those

 

[Blumlein 88]

If you can find one! The 7520s and their HiFi counterpart Z1000 have become rarities.

Edit: As I say that, I find some on a well known auction site. Interesting. I might have to pull the trigger on those

Makes me wish I had splurged on some 7520s then. I've got 7510s.

 

[Mad_Economist]

Very true, but more amenable to mostly solving problems below Schroeder frequencies. There is always get a bigger room or listen outside.

What, and deny myself all those sonic picture painting delayed reflections? Heretic!

Oh okay, so no solution, just a literal limit of reality of the matter then, correct?

Essentially yes. I suppose in premise if you had some consistent way of doing a correction of headphone output based on eardrum sound power, that could work, but I have no idea how you'd do this.

So the "problem" gets worse with smaller structures, or what's going on here? Would it then be safe to assume IEM's are the worst offenders (meaning introduce the worst examples of variability above 10K), or do they recoup a bit considering they bypass notions of cancellations and such other things that I assume would be more prevailent within the space of a headphone cup?

IEMs don't really have positional variation other than the resonance of the volume between them and the driver - however in some respects this is trading many small problems for one large problem.

 

[Francis Vaughan]

IEMs don't really have positional variation other than the resonance of the volume between them and the driver - however in some respects this is trading many small problems for one large problem.

Siegfried Linkwitz created explicit compensation for this problem addressing a couple of IEMs.
https://www.linkwitzlab.com/reference_earphones.htm
He notes that the precise frequency of the frequency is dependant on the individual users physiology.
I have always intended trying out his solution with my ER-4S phones, but still have not done so.

In principle, since we are addressing a single simple resonance, the problem can be exactly neutralised by crafting a filter section with the right parameters. The problem is that you need a per-user set parameters.

 

[Jazmanaut]

One issue I am finding with the pinnae on 5128 is its material characteristics. It is soft and flexible which is its claim to fame. But it also has high stiction with similar material such as the silicone tips of IEMs. They are like magnets and instantly stick to each other, making it harder to find the ear canal and push it in there. Hard pinnae don't have this issue. Real ear has a bit of this and the reason I like silicon tips but not as much as what I am seeing on 5128.

I guess the solution is to use foam tips.

But I also worry whether the pinnae is grabbing onto the pads of circumaural headphones. Hard to know how bad this issue is as the pinnae is occluded by the headphone pads.

I had absolutely best and most repetitive results with memoryfoam tips.

 

[Hipper]

Listen to speakers.

One problem, that's not a solution for headphones

I've often wondered whether you could listen to stereo speakers placed in such a way that you are in the centre of a straight line between the two speakers - effectively like big headphones but a little further away so that the multiple drivers can integrate. If things work correctly you should still be able to create a phantom image.

I realise that our ears have evolved to hear mostly what's in front of us - or perhaps it is our brain that has learnt this.

 

[MayaTlab]

In principle, since we are addressing a single simple resonance, the problem can be exactly neutralised by crafting a filter section with the right parameters. The problem is that you need a per-user set parameters.

ANC IEMs like the AirPods Pro feature a microphone directed towards the ear canal. With continuous improvements to mic quality and real-time DSPs, would it theoretically be possible with this in-ear format to automatically reach a FR target within 1-2 db across the whole audible FR range, including elimination of the ear canal resonance, or would there still remain "laws of physics" limitations that would require a different approach to overcome that problem ?

With over-ear headphones the pinna poses a different challenge, right ? Could real-time analysis of the 3D shape of everyone's ears, maybe with a technology akin to Apple's Face ID, associated with real-time 3D calculation of the user's pinae effect on sound, theoretically be a solution to that problem ? Or would it be possible to already reach decent results with a multi-mic system inside the cup (I believe that AKG uses two mics for its one-time - not real-time - TrueTone evaluation) ?

 

[Bob-23]

I've often wondered whether you could listen to stereo speakers placed in such a way that you are in the centre of a straight line between the two speakers - effectively like big headphones but a little further away so that the multiple drivers can integrate. If things work correctly you should still be able to create a phantom image.

I realise that our ears have evolved to hear mostly what's in front of us - or perhaps it is our brain that has learnt this.

If the recordings were mixed with speakers in that position, it should work.

 

[Balle Clorin]

Not exactly what you asked for, but the comparison of the flat-at-the-listening-position-with-an-omni-mic Revel F208s in the Harman reference room
View attachment 78685
to the response at the mannequin's eardrum (black line) under the same circumstances (after averaging of a few angles and smoothing):
View attachment 78684

Many thanks, so the flat in room from the top plot is transformed to the bottom after passing the outer ear and ear canal= this is what the in-ear phones should give i guess, maybe the ear-canal after the in-ear gove som addional natural eq?
When I measure my ear-plugs directly onto the mic on table , it looks somewhat like the lower curve.

 

[FeddyLost]

IEMs don't really have positional variation other than the resonance of the volume between them and the driver - however in some respects this is trading many small problems for one large problem.

So, the custom mold is the solution. Maybe. Anyway, shelf equalization of bass is easier than equalization of some crazy unpredictable sawtooth in mid-highs.

 

[Scgorg]

Many thanks, so the flat in room from the top plot is transformed to the bottom after passing the outer ear and ear canal= this is what the in-ear phones should give i guess, maybe the ear-canal after the in-ear gove som addional natural eq?
When I measure my ear-plugs directly onto the mic on table , it looks somewhat like the lower curve.
View attachment 78723

Without a properly sealed front volume you won't get any reasonable measurements, which is pretty clear from this plot. Your measurements are not measuring what the actual IEM response looks like.
First, the low frequencies (in this case "low frequencies" means everything below 1KHz) look like they are simply noise, which makes sense considering the fact that the driver won't really have any output at those frequencies.
Second, the high frequency behaviour is inherently not accurate either due to the lack of an ear canal resonance with this measurement. This is the answer to your question about the ear canal giving "some form of EQ". The ear canal resonance is a quarter-wavelength resonance that happens at a frequency related to the internal volume of the canal ("tube").

 

[PeteL]

Listen to speakers.

I guess a point could be made that most music is mixed and mastered on speakers, using headphones just for reality check or to pinpoint micro adjustments, and then subsequently the scientific approach of Olive and others craft their analysis based on speaker listening as reference. Now Is there a real validated science based reason for this or it's mainly historical? That's an interesting debate, we can point the issues of audio reproduction where the dimensions of the acoustic environment are significant compared to High frequencies wavelenghts, but, we can't deny that the Headphone/earphone market, despite that, is growing at a much higher velocity than the traditional way of speaker listening. So wondering if this down the line could lead to a paradigm shift, like, would the music actually sound different if it was mixed for headphone first, and for speaker as a side tought? I'm just being philosophical here, sorry If I derive the subject.

 

[Tks]

What, and deny myself all those sonic picture painting delayed reflections? Heretic!



Essentially yes. I suppose in premise if you had some consistent way of doing a correction of headphone output based on eardrum sound power, that could work, but I have no idea how you'd do this.


IEMs don't really have positional variation other than the resonance of the volume between them and the driver - however in some respects this is trading many small problems for one large problem.

And just very quickly as to not take up anymore time for serious discussion. If you were tasked with furthering either IEMs or headphones (taking into account knowing you're working on headphones now). With the issues we just spoke about where you trade smaller numerous issues, for one large one. Which do you think in theory would be worth perusing first (one where you would make the most headway in rectifying). I'm having trouble grasping as to where more headway could be made given the current position with respect to the issues of each listening medium.

I feel even with all issues IEM's have, something like like Etymotic with deep insertion, and the substantial isolation, and proper recordings, I feel IEM's produce some sounds that fool me into thinking it was something in my room (ambient recordings and such sound hauntingly close to as much fidelity as I could ever need). Toss in binaural recording, and we've got something that could fool a person if he weren't wearing something the equivalent of earplugs that the Etymotics are (tbh, even normal universal IEM's outside of the Etymotic offerings achieve this with things like foam tips).

Speaking of eartips though, really wish companies would stop (in the same way I wish would stop with headphones), the production of completely round tips, and make them more oval (likewise with headphone earpads - where I think Sennheiser really shines, especially on their HD800 line I wish someone would stop messing around and emulate such spacious interior that I assume is partly responsible for is producing all the "amazing soundstage" anecdotes).

 

[Blumlein 88]

I guess a point could be made that most music is mixed and mastered on speakers, using headphones just for reality check or to pinpoint micro adjustments, and then subsequently the scientific approach of Olive and others craft their analysis based on speaker listening as reference. Now Is there a real validated science based reason for this or it's mainly historical? That's an interesting debate, we can point the issues of audio reproduction where the dimensions of the acoustic environment are significant compared to High frequencies wavelenghts, but, we can't deny that the Headphone/earphone market, despite that, is growing at a much higher velocity than the traditional way of speaker listening. So wondering if this down the line could lead to a paradigm shift, like, would the music actually sound different if it was mixed for headphone first, and for speaker as a side tought? I'm just being philosophical here, sorry If I derive the subject.

Not going to drag this off topic, but yes mixed for headphones first would be different.

 

[outerspace]

Hi, Amir! Sean Olive mention experiment when they made headphones recording from ear simulator and then compare them with just equalized recordings. They compared all differences besides FR. "People couldn't really tell the difference", he said. There is a even better version of experiment, I think. You can made headphones recording with equalized output to made precisely flat recording and then compare this recording with non-modifed original track. If you interested and if you want I can support this experiment with recordings, precise equalization, alignment software, technical suggestions etc. Cuz I'm already made something like this.

 

[Balle Clorin]

With a little EQ in Roon my Sony MH755 bought for 6usd sound just fine , and I canceled all plans to buy headphones. Would be interesting to have something so cheap measured for low end reference.

 

[FeddyLost]

So wondering if this down the line could lead to a paradigm shift, like, would the music actually sound different if it was mixed for headphone first, and for speaker as a side tought?

For sure it will sound different.
You can find some complains in sound engineers' forums that more and more current customers need their products to "sound good" not only in IEMs but in 1 pair of IEMs for 2 people and from portable battery speakers or cranked up phone...
So, all the mix is different.