B&K Curve with In-Ear Mic (Experimental)

TBH, I preferred my in-ear measurements without the B&K target. Flat with the in-ear measurements seemed to be quite close to speakers with a B&K curve at the listening position.
Still, this will depend on your mic position and calibration. In this case I think I'd agree more with your previous measurements.

 

Also wanted to add that the B&K curve at the listening position measured with an omni mic doesn't sound the same for every speaker setup. In other words: The same measured FR measured with traditional methods will not get you even close to the same FR at the eardrum, because the head is not an omni mic. There are some other important factors like angle towards your head, direct vs reflected energy, speaker directivity, ...
What I'm really saying with this is that a comparison between speaker FR and headphone FR isn't easy. This is also why I take measurements at the listening position, but also take measurements at about 1m distance for speakers. When I'm bored I will take measurements at the ear canal opening, just like I do for headphones. Music also isn't only continuous tones and I feel the FR of the direct energy is more important than the FR of the reflected energy for speakers. This is why I think the speaker FR targets like the B&K target are only a starting point. In the end it's best to use the "this shit sounds good" curve.

For the in-ear measurements I wouldn't use a special target curve. Maybe a compensation for the upper treble rolloff. For the coupler measurements I think a target is needed, but I also believe it's just best to let people figure out their own targets based on what they hear. What people shouldn't forget is that headphones where the drivers are bigger and at a larger distance from your ears will need a different target than headphones like the HD6X0 or even on-ears.

Also, as you noticed, some headphones just have weird measurements, like the Elear with the in-ear method. The Elear and Utopia treble does measure much better on the open foam coupler which I think is closer to what I heard.

 

LOL, what I wanted to say is that it's just best to listen and not to measure sometimes. Anechoic measurements can be very useful, but I think to get a good picture it's necessary to measure at more than 1 angle. Probably better to measure at 72+ angles. When I measured my speakers in my backyard I ended up getting reflections from the ground, mainly because I voice my speakers for 2m listening distance. I think at about 2ft from the drivers I got reasonably close to anechoic conditions over most of the audio band, but also have to measure in the far field. I ended up placing some absorptive material on the grass, but once you go low enough in frequency it won't do much, either. Even if it's 1ft thick, or thicker. Also, in cold, rainy Germany it's often not practical to measure outdoors. And the speakers are heavy.
I also think that for speakers the room has to be part of the equation.
In the end I don't worry as much about the off-axis FR as I should. My speakers have horrible vertical off-axis FR because of the wonky 1st order filter and they sound super shouty towards the back.

Also, the room interactions is where the omni mic will differ most from what you hear. For the anechoic measurements you only have your speakers as a sound source. For the home environment you have all sort of reflections which will all have a different tone depending on the angle, based on the HRTF. The omni mic won't capture these interactions. This is for example why certain rooms and certain speakers sound better with a 2-3kHz BBC dip and others don't (IMO). In the end I would rather tune my room to have less reverb in this region than to voice my speakers to have less output there, but those two approaches will of course sound very different.

 

The Low 30's seem to be around where my 650's drop off, but I've only exclusively done the Dynamat/Damplifier mods which may skew things slightly. Many thanks for all the measurements done, I've been lurking for too long.

 

With the modded HD650s I generally found the 1-5kHz region to be slightly above the 500Hz level, which was a little off-putting to me. Also, there sometimes seems to be a slight dip at 2kHz which I didn't like. I can't say how yours sounds. I'm pretty sure you used some magic to bring the upper mid level closer to the fundamentals region.
The 5kHz bump is weird in that the HD800 has a very similar 5kHz bump on my in-ear measurements, but I don't hear it as much. The HD6X0 drivers have a resonance around 5kHz, while the HD800 seems to excite a natural resonance that my ears have. (The HD650 drivers possibly have less of this than the HD600. I think it's mostly related to the rear magnet opening hole, as some back damping can get rid of it for the most part.) Even if the result looks similar in the measurements, it doesn't sound similar. I do hear a slight 5kHz bump (+ resonance) on the HD6X0s that I have heard, but I also generally hear a big 6kHz+ dip, which sort of masks it. And of course the bass helps, too.

I'd say your bass extension is a little too good to be true. I think the 40Hz level should be in line with 1000Hz and 20Hz should be 5db below that. Bass hump should be about 3-5db. Too much for me. I think it looks less than ideal in this graph because of the B&K target compensation.