Well, about 'real sounding music' part... It's really hard to define 'real sounding music' because it is not just instruments and/or human voices that determine the sound.... The environment takes a huge role on how sound is produced. I mean, it's obvious that sound signature of a large orchestra hall is quite different from that of a small, enclosed room or inside of a church. Real sound music should be reference, but in reality it is also completely impossible at the same time.

Indeed, and then there is also something like how a recording was mic'ed.
How loud the SPL is during an evaluation obviously also has a great impact in how 'neutral' a HP is perceived.


Sure, all instruments sound quite different in different 'rooms' and/or outside.
Yet, in all these circumstances we can pick out the 'real' sound of the instruments which, I think, has to do with the first wavefront that our brain finds most interesting for determining 'what' it is we are hearing and where it originates from (localization).
Amongst it by comparing incoming sounds to our own (perhaps often misguided) reference of how something should sound.
 
The (later arriving) reflections tells us something about room dimensions/conditions.
When you measure the FR of the first wavefront and add room resonances + reflections the total SPL is not 'flat' in FR while the source could be.
Yet, the 'sonic signature' of the sounds that arrive first may well determine how 'real' something sounds and coloration is perceived as 'surrounding' and not necesarilly as a coloration of the source.

A well made recording in those different recording circumstances should still sound 'real' and the reflections should still be perceived as an image of the conditions/surrounding it was made in. At least it does to me on better headphones.
The 'speaker in a room emulating' headphones don't seem to do it for me (too dark) but may well be a 'reference' to the armchair music appreciaters.

Of course 'training' with real (unamplified) music remains essential if you want your built-in reference in your head to work better in an absolute way.

I think the brain calibrates itself continuously though.
When I start listening to a KSC75 it sounds quite 'real' to me after a very short time already.
Yet, when I listen to a better headphone or real music first the flaws of the KSC75 are immediatly appearant and it takes a LOT longer for me to 'believe' the KSC again.

remains an interesting topic... what the reference should be and how to determine it.

In the end I am convinced it is really all moot and think that what's one man's reference is another mans nightmare.
Everyone should determine for themselves what sounds good to THEM, not what someone else feels is 'correct'.

Well, about 'real sounding music' part... It's really hard to define 'real sounding music' because it is not just instruments and/or human voices that determine the sound.... The environment takes a huge role on how sound is produced. I mean, it's obvious that sound signature of a large orchestra hall is quite different from that of a small, enclosed room or inside of a church. Real sound music should be reference, but in reality it is also completely impossible at the same time.

Not really. In most cases instruments are mic'd at the source so there is very little environmental concern, at least in how you are describing it. Acoustic treatments are usually done to prevent the environment from interacting with a recording rather than trying to capture the ambiance of a specific venue and chasing each others tails. Granted, there are recordings and live performances where that is part of the equation (hearing specific reflections and ambient cues), but they are the exception and not the rule. So they shouldn't play a part in chasing reference gear or masters unless you know those particular colorations intimately and can compensate. That's actually true of any recording as 99.989% of recordings have something off about them somewhere, it's simply a matter of degrees.

Point is, for reference, audio engineers would do what they can to remove the environment from the equation. As you hear better gear (sources, mics, etc.) and better mastering, you realize just how good and transparent quality recordings can be. As a result, I feel totally confident when listening to gear that I can pretty much tell exactly where in the chain something is off without having to confuse myself chasing sonic phantoms. We actually had an interesting incident of just that when performing some ADC rips a week ago. Having enough experience with our gear, we could use deduction and pinpoint the source of a specific issue we were hearing in the recordings.

I suppose what I'm saying is, it's a lot easier to achieve a reliable level of accuracy even if it may be very difficult to achieve extreme precision. So don't worry about being precise, worry about being accurate.

I think this is similar to what solderdude was getting at. Sorry, I was typing too slowly while watching Scarface.

So how to determine reference? Well, I like to go off just normal daily ambient conditions just as you would attending a live performance or playing an instrument. Unless you work with jackhammers or on a tarmac with boeing 747s, your ears are largely set for day to day activities in my estimation. So it's okay to start there. Granted early in the morning versus later in the afternoon might see a normal rise in SPL compensation, but tone and timbre would remain relatively consistent within Fletcher Munson and the fundamental level at which live instruments are performed is relatively consistent as well. Now some prefer to listen a lot lower for various reasons and preferences too. At that point, different gear could very well provide more accurate sound at those levels than say gear that would perform better at higher SPLs. To me, SPL is the only place where there is a very real audible sliding scale happening. Personal references and listening conditions are very rarely mentioned in audio reviews, yet are some of the most crucial.

Yup.

I'm just excited for the day that science catches up to Luis' or OJ's ears.

I'm just excited for the day that science catches up to Luis' or OJ's ears.

Already has to a large extent. It's called Changstar and why we don't worry about HRTF, diffuse-field, free-field, reality-distortion-fields, Sada-Flores-Chen no-compensation, or whatever other science blah blah blah that over-thinks or over-analyzes things that don't need to be over analyzed. 90% of the people who talk about this stuff really have no clue what they are talking about.