"if I understand correctly, you're problem seems to be that headphones (and even monitoring or high-end headphones) that sound flat are not what you're after."
Yes. At least I seem to thinks so (a working hypothesis).
I need a sound source that is as neutral as possible as a sound source. This means NOT taking into account (at least not A LOT) generic head and outer ear related transfer functions.
It doesn't need to SOUND neutral to the test subject (it hardly will), but it needs to produce a neutral response from a test signal BEFORE that sound enters the ear canal (or even pinna or reflects from head).
Of course this means taking into account some of the head & headphone coupling related sound colorations, but it is complicated not to overdo it as pinna is part of this equation and VERY personal (i.e. hard to make generic functions about it).
If you look at the field of psychoacoustics a lot of it seems to concerned with what happens with certain kind of excitation if that excitation exists at near the tympanic membrane.
Hence, these tests usually take into the head/ear related non-linearities and try to compensate for them IN ORDER to produce a desired kind of pressure signal at the tympanic membrane.
I don't want this. I want the ear / head to distort the signal the way it naturally does (differing with each test subject). Just like with loudspeakers.
Before you yell "but loudspeaker are different", I will agree with you.
The problem is finding a sound source that is NOT a loudspeaker (I don't need to take the room into account), but is not either a generic HTRF-function tweaked headphone that tries to sound neutral to all ears (and probably fails on quite a few accounts).
"You want something that measures flat. But virtually all headphone designs take head-related transfer functions (HRTF) into account and try to equalize the sound in some way."
Indeed they do. I agree.
I understand that headphones ARE a special case (especially as they form a semi-close or close cup around the ears and as such can cause a lot of distortion themselves).
However, often this tweaking of the headphones sound is taken so far as to neutralize the effects on the pinna and even inside the ear canal, which to me would make the headphone useless for my purposes.
It may sound neutral to some heads and ears, but very hard for me to use, unless I knew what were the HTRF functions and tweaks used.
"But since audiologists aren't too concerned about hearing loss above 15kHz, audiometric headphones will not give you the treble extension you're looking for, up to 30 or 40 kHz."
Yes, I've just learned this. Thanks
JaZZ: "I'm still not quite sure what you want. There is no headphone frequency response independent of the ear and its necessarily individual shape/non-linearities."
Indeed there is not. This is also what I would like to do some testing about.
However, if the headphones I use ALREADY takes into account some statistically derived generic HTRF function (and does additional tweaking on top of that), for which I have know mathematical model...
Then how can I test for the non-linear effects of the head/ear combination if I don't even know how much the headphone already tweaks my input signal, in order to try and avoid some of the non-linearities of head/ear combo?
"Any measurements through an artificial ear (weighted, if need be) come closer to the hearing impression than without it – whatever the measurement conditions thereby may be. "
I wouldn't want to disagree with that. At least not yet, if ever
I know what I'm asking may be impossible, even stupid to some, but it's very hard to quantify that notion scientifically, if I can't test it. Based on preliminary literature research this has not been properly tested yet.
"What's the exact purpose of this? You had ultrasound on your wish list, but then you're considering yourself some «audiometric» headphones which barely reach 12 kHz... "
Well, my knowledge in audiometric headphones is (was) even less stellar than my knowledge in headphones in general. Thanks for bringing me up to speed
But what I would like to test for is in particular the non-linear functions of head / outer ear (especially) with as neutral sound source (that is not a loudspeaker) as possible.
"well as extension. Why should now a HP 1000 go for an ideal audiometric sound source and not any Stax for the same price? "
Because STax cause 2x the price the best dynamic headpones I can get WHERE I live (of course I can always go through the import hassle, but after shipping charges, insurance, 2-5% custom fee and 22% of value added tax added - ALL CUMULATIVELY on top of each... well I'm back to normal Stax range).
Maybe the cheapest stax could do it price and performancewise, I'm not sure. I can't seem to find reliable and comforting measurement data for them, however.
Also, correct me if I'm wrong, but I lived under the impression that most Stax phones for which accurate measurement data exists produce: very irregular frequency response, very sharply declining high frequencies and even some kind of intermodulation distortion from the connection frame?
I've only seen HE60 been supported and considered highly by audiometric/psychoacoustic researchers and that headphone is way out of my price range.
Please understand that what I write above is a serious intention to learn more about my choices and I mean no harm.
I appreciate all comments so far and would like to hear more, if somebody has any ideas.
English is my not my native tongue, so if somebody feels like I'm offending them, rest assured that it is not my intention. Eloquent expressions just don't come naturally to me - at least not in English that is