What James is saying is basically correct. The research was conducted by Toole. Numerous studies have supported that view. While you can look up the various AES published articles, it’s probably easier to buy Tooles books and read those. They do a better job explaining it in lay terms.
I own the Acoustics and Psycoacustics of Speakers in Rooms, and have had brief interactions with the man himself (he posts on another forum, or used to at least). I've also done the bulk of the hearing training regimen he created as part of his studies.
The take-away I have from double-blind testing he conducted at HK was that the major issue with a negative effect was non-flat off-axis performance. I have even experimented with [purchased] speakers built around that premise by someone perhaps well described as an true believer in that regard
None of this has anything to do with how the instrument is recorded. It has to do with how we perceive speakers in a room as compared to outside. A speaker that measura flat anechoicly will have a tilt in the response of about 1db per octave from 20khz to 20hz.
You seem to be on a different subject. This is not about the speaker's performance. This is the statement I responded to
"Very few people would enjoy a system that measures flat in room at the listening position. "
Can you point me at something that indicates that a non-flat performance is preferable to one which exactly reproduces the source material at the listening position?
I also wonder if the anecdotal experience that might support that isn't training.
I know a lot of people who dislike classical recordings where the mics are near the instruments and then the orchestra is placed in mixing. They much prefer stereo (usually with head-shaped baffels) in listening positions in performance halls. I am quite the opposite. I often think that my experience being *in* the pit when the music was being performed compared to most other's experience being in the audience creates an expectation of what is "better".
Further: where I do listen to recordings with strong room influences (such as in a cathedral); I must prefer headphones or a heavily deadened room. I suspect I would very much like an anechoic chamber for listening; though I would not be surprised if flat would be an aquired taste. Too much time with non-flat changes the pallet.
It’s also important to remember that the standard approach or taking in room measurements is not a very intelligent method. In the words of Toole, it’s a “dumb” method. That is, it’s just measuring the steadystate based on omnidirectional pickup at a single location.
This feels a bit like a straw-man. You are arguing the conclusion by simply rejecting the premise.
As such, applying eq above a certain point means changing a response based on bad data. Those measurements do not reflect what the ear is hearing. It is very likely you are doing more harm than good.
This seems to tie into your earlier post rather than the one you are responding to here.
But I'm worried that you've just changed the goalposts.
You had "Nope! That [frquencies above 500Hz] is not what you hear "
You've said that EQing above 500Hz is bad, which I've asked for the reasoning for.
As I understand your reasoning in your more recent post, it's because "people take bad measurements". Is that an accurate understanding?
Indeed: didn't Toole's DB testing show in many ways the exact opposite? That speakers putting (measured) flat responses off axis were preferred over speakers which put (measured) non-flat off-axis response? How is this not at least a partial validation of a relationship between measured response and subjective listening preferences?
There's a vast gulf between "most people EQ wrong" and "EQing is bad".
PEQ and normal measurements using normal processing like REW does cannot be used to do in-room full bandwidth eq. Toole, Geddes, and many others like them would further argue that no tools can do this and Dirac and Audyssey should not be used.
Has he actually made that argument? Has he actually stated that no tool can measure sound usefully?
Did he do blind studies I'm unaware of that said the results of such EQ were universally bad?
However, I have read all ir Diracs published papers in their algorithm and believe that it has merit. Their approach uses a number of spatially varied measurements to detect those aspects of the response that are caused by reflections and diffraction. It can then ignore those and look for those anomalies in the response that are spatially constant. There is a little more to how this happens, but the fact remains, it then applies EQ in a manner that can avoid those problems. Because of its use of mixed phase filters, you end up with essentially an inverse transfer function of the response accounting for those aspects of the response that you shouldn’t eq.
I fear this paragraph includes things I have too little background with to agree or disagree.
