Fidelity means sound A is indistinguishable from sound B. And as far as I'm concerned sound is a fluctuation of air pressure. A thing with measurable characteristics regardless of how it is analyzed by humans, animals, aliens etc.

Are you 14?

Again the same question arises. Who shows the true difference between 990 edition and 990 pro. Headphoneinfo or the others. Are measurements trustworthy?

Only the site who has done measurements of the 2 different models on the exact same rig can show the DIFFERENCES between the 2 models.

On the question are measurements trustworthy than I would have to say NO.

For instance, have a look at Tyll's raw plots (the grey lines under the 'final' measurement).
They are taken on the same rig but with different positions on the rig.
These different positions, more often than not produce WILDLY varying plots.
Same headphone, same rig... which of the grey lines would you say was accurate in an absolute sense and what correction would you use, OW or Tyll's or another ?
I get similar on my rig and can get VERY different (as in 10dB even) different plots depending on the position of the headphone ON the rig.

Most often I take the one that represents what I hear closest in a position it would have on a real head.

Even this isn't a guarantee, take the large diameter of the pads for instance. They may break seal on some heads, and not of others, hears may do the same (don't have much !)
Then there is 'correction, compensation, HRTF theories' etc.
All of these will produce different absolute results compared to perceived sound.

Having said that ... all measurements out there may give you a pretty good idea of how they sound TONALLY.
All DT990 measurements will show roll-off in the sublows, a peak in the lows, recessed mids and a substantial peak in the highs.
Well not in the plots you mentioned, they stop at 100Hz.
low and treble roll-off, peaks, dips and general tonal aspects all show how the tonal balance is and that's the 'value' of measurements.
They need to be accurate in an absolute sense and to your ears/taste if you want to base EQ on them.

Then there is spread between drivers even of the same MODEL which can be several dB's and shouldn't be understimated.
Resonances that may occur can differ substantially from rig to rig.. Pinna or no pinna, hard surface or soft/porous, flat or headshaped surface even temperature van make a difference in seal when having memory foam in pads.

Too much variables ... but ... if you found a website (or made your own rig) that you trust you should stick to them.
Pirates often can find themselves in plots made by Marv, UB or Hans.
If you feel other plots are more 'correct' to your perception than you just use those and ignore, or take for granted, the plots shown here.

Only when you have made measurements yourself you start to get a feel how difficult it is, even more so in the speaker realm with reflections and resonances etc.

That's called freedom of choice..
I make my own choices, so can every one else.

Here (in Pirate land) plots are viewed as indicative and to give an idea of how it measures compared to other measured plots, not as absolute truth.
No one here has ever claimed absolute accuracy.
Take the plots for what they are ....


Back on topic .... I really like the HD650 ... EQ'ed flat... not in stock form.

Considering the fact that radio shack spl meters are inaccurate above 2Khz and tend to overestimate SPL, HD 650 frequency response might resemble a flat line with a downward slope above 2Khz when measured on a simple plate as a CD.
The problem with this kind of measurements-I believe-is that they do not take into account the distance between the transducer and a real ear as well as the dispersion characteristics of the former.
I would be glad to have your opinion on this one. Can simple measurements on a plate-without the use of dummy ears-accurately capture the FR differences between different sets? Only their relative differences. Not how their sound is perceived.

Written below is my personal take on this and NOT proven science !
many may disagree on my views which is O.K.

Radio Shack meters cannot be used for accurately measuring FR, not even in 'C weighing' and 'fast' mode.
The moment you apply a baffle around a microphone tip the FR changes considerably.
The tip of measurement microphones isn't very small in diameter with a large 'stick' behind it for nothing.
Reflections a baffle creates must be avoided when measuring.
You can only use these meters for measuring average SPL in free air.
The CD also reflects sound quite differently than skin.

Once the sounds reach our outer ear what happens inside doesn't matter AND has to be corrected anyway so why use a tube inside to mimic an eardrum.
The ear drum itself 'hears' a completely different sound than what's outside of the ear anyway.
I think what's OUTSIDE of the ear should be 'flat' and not 'speaker like in a room' but others may seek 'speaker like in a room' and detest really flat.
preference.... POV... religion ... whatever rocks your pirateship.
HP drivers pump soundwaves directly into our ear canals and the Pinna/Concha only marginally affects what goes in there.
With speakers the Concha and Pinna have a FAR greater and different effect as the sound comes from another direction all together and thus 'compensation' of HATS need to differ as well.
HOW the soundwaves are affected depends on driver diameter, position and air volume as well as (freq dependent) seal of pads.
I don't think HATS used for evaluating speakers is perfect for measuring headphones either.
Every type of rig has its own + and - assets, no rig is correct in an absolute sense. Some are closer than others.

I have found in experiments that the absence of a Pinna can make a substantial difference in measurements with HP-A yet very little difference with HP-B even when both are over-ear headphones.
No way to tell what is accurate in an absolute sense. One can only guess based on what they hear and 'adjust' the rig accordingly if needed.
Myself I only use a form of 'adapter' when measuring on-ear and don't use a Pinna for several reasons.
I don't like in-ears at all, nor their sound, and think these are nigh on impossible to measure accurately compared to how they are perceiced for a number of reasons, so stay away from that trap.
There SHOULD be a standard for this.

Would you like to share your settings in REW filter settings form?
I mean like this:
Filter  1: ON  PK       Fc    50,0 Hz  Gain -10,0 dB  Q  2,50
Filter  2: ON  Modal    Fc     100H z  Gain   3,0 dB  Q  5,41  T60 target   100 ms
Filter  3: ON  LP       Fc   8.000 Hz
Thank you in advance

Can't do that.... not because I feel it is proprietary or secret or don't want to share but because I don't use any EQ in REW.
I use my own rig (its described here and there) with a WM61A capsule and a home made mic pre-amp with  analog (passive and active) compensation for the mic capsule.
The line signal into my EMU0404 is thus already compensated (and hisses a lot so cannot do distortion tests).
The only referencing I do is my EMU- amplifier (Ember proto) - EMU line loop.

My headphones are all compensated in the analog plane and those are all well documented. I don't like digital EQ.

My observations and views generally jive with Solderdude's

Pinna makes a difference with certain kinds of headphones, but not all. I ended up with a simplified pinna for the purpose of being able to take measurements of supra-aural headphones. Distance on microphone plane of solid-coupler to the headphone driver will make a difference to the bass-treble balance of the resulting measurement. The can be tweaked. I prefer flattish instead of speaker-in-a-room response, hence no "DF" or "FF", unless IEMs, which I feel a little bit of that kind of response.

As for the plate or the baffle, some of its effects have to be considered. The headphones lie on skin and bone, which is essentially a baffle, but with different characteristics from a CD. You can experiment and try lining different materials on top of the CD. Also, correct seal with a baffle is required to get good results on bass. Too much seal overemphasizes bass measurements, hence I implemented hair simulator strips. A free-air or semi-free air measurement provides useful information as well, I think our ears can filter out even the early reflections in the cup - but this method has issues sometimes with seal / pressure to have the correct bass measurement.

Certain measurement techniques work with certain kinds of headphone designs better than others. Eventually, I had to pick one that worked best on the majority for the sake of consistency.

As far as HATs using to measure speakers, that doesn't make any sense unless you are trying to develop complex transfer functions exploring the inner workings of the ear and torso. Speaker measurement is always a mic wand at listening position in the studio or mic wand at 1 or 2 meters out at tweeter level by speaker builders. With a HATS, by the time you are done applying the compensation for pinna, concha, ear canal, it's supposed to be the same as the mic wand.

HATs do have different directivity characteristics than a wand mic(s), and tend thus lend themselves well to stereo recording. In any event a measurement must be taken with a wand mic and compensation must to applied to the HATS.

I don't mind giving hints, but I generally don't talk about or give specifics on my measurement rig(s) or techniques. Several cases where people who have seen my rig have gotten ideas / asked a bunch of questions and started to implement their own measurement systems to sell headphones. I'm sure all of these people could have developed a system themselves, but I don't like making it any easier for them.

Again, I think Dr. Olive is a better source for such discussions. I treat this as a hobby trying to get quick-n-dirty results and crude insights. I don't claim accuracy nor intend for the work here to be in academic papers or journals.

I have found in experiments that the absence of a Pinna can make a substantial difference in measurements with HP-A yet very little difference with HP-B even when both are over-ear headphones.
No way to tell what is accurate in an absolute sense. One can only guess based on what they hear and 'adjust' the rig accordingly if needed.

I wonder if there's a correlation to the driver size/wavefront and the interaction with the pinna etc.

Hmm now my brain is pondering how to create a "microphone" that can map a 2d space rather than point.