I am taking a fresh look at the HD 800S's graphs and frequency response, using a slightly different (and hopefully easier and better) approach with Jude's HBK 5128 measurements.
The approach I use is similar to Headphones.com's target model. But a bit different. Instead of using Oratory1990's 5128 diffuse field curve, I use HBK's 5128 DF measurement for compensation. And compare the results to the
sound power responses of good loudspeakers... or to a
slope in the same range (roughly -1.0 to -1.5 dB/octave). I refer to this as the "DF+SP model". The objective of this model is to
approximate the response of neutral loudspeakers in a good room at the DRP (eardrum reference point).
HBK's DF curve only has a 1/3 octave resolution though, which limits the results to 1/3 octave rez. This is something that was pointed out to me by Oratory1990 on one of the Discord measurement channels. And something I neglected to take into account in some of my earlier tests with this approach here...
https://www.head-fi.org/threads/sennheiser-hd800s-measurements-brüel-kjær-5128.951181/
...But I think it might make some things a bit simpler and easier.
The first step in the process is to convert Jude's raw 5128 HD 800S measurement to the same 1/3 octave resolution as HBK's DF curve. This is easy to do in Equalizer APO's Configuration Editor by simply converting a Variable GEQ curve to a 31-band GEQ curve. The net result looks like this....
After compensation with the inverse of HBK's 1/3 octave DF curve, the HD 800S response curve looks like this...
This curve can then be compared to the sound power responses of good loudspeakers (Pierre Aubert's
Spinorama site is a good source for such sound power and other speaker data), or to a slope in the -1.0 to -1.5 dB/octave range...
(blue line = -1.25 dB/octave slope)
HBK's 5128 DF curve is only accurate up to about 15 kHz though imo. And the compensated FR of most good headphones will drop off above that, like the HD 800S does above.
Although it is not a perfect slope, the compensated HD 800S curve above is not that far away from the sound power responses of some good loudspeakers, which will often have a dip somewhere around 2 kHz, where the midrange and tweeter drivers cross over. (Some folks derogatorily refer to this as the "BBC dip".)
There are quite a few newer speakers with sound power responses closer to a linear slope like the one shown in blue above though. And that slope is generally in the -1.0 to -1.5 dB/octave range. So that's why I look at both types of response: the actual SP measurements of various good loudspeakers (especially those used for mastering/listening to music), and also a linear slope in the above range, for comparison.
When compared to a slope, the HD 800S looks a bit too depressed in the upper mids. And also a bit too bright in the 6 to 8 kHz range of the treble. A bit more lift and extension into the bass/sub-bass frequencies would probably also be more desirable.
Even some relatively inexpensive nearfield studio monitors, like JBL's 306P MkII and Kali's LP-6v2, can extend down pretty close to 40 Hz in the sub-bass before beginning to roll off, for example. (And can also be paired with a sub-woofer, for an even better sub-bass response.) You can get an idea of what the sound power response of these two speakers looks like here, for comparison...
https://www.spinorama.org/speakers/JBL 306P Mark ii/ASR/index_asr.html
https://www.spinorama.org/speakers/Kali LP-6v2/ErinsAudioCorner/index_eac.html
The JBL has a fairly sizable dip in it's sound power response at about 1.5 kHz, where the tweeter and midrange drivers cross over, which is not that unusual. While the Kali generally approximates a slope in the -1.0 to -1.5 dB/octave range.