That is annoying. Any idea what EQ program Oratory1990 is using?
Edit: Are you sure? The graphs look pretty similar to me, except that I reduced the low shelf a little to taste.
Most of the CSD plots of headphones I've seen only go out to five milliseconds. Room acoustic measurements generally look at hundreds of milliseconds, so it's hard to see how a couple milliseconds would have much effect. Headphones usually have much lower distortion than loudspeakers as well...
It looks like the AutoEQ values actually roll off high frequencies quite a bit more than Oratory1990's PDF:
Oratory1990 doesn't try to correct anything above 10kHz or so:
(I subtracted about 1.5dB from bass in both examples to compensate for GoodHertz CanOpener crossfeed plugin.)
It looks like oratory1990 measurements are included in the AutoEQ output. Are you referring to the referenceaudioanalyzer category?
Edit: Actually, it looks like Oratory1990's settings are a bit different. There's an index here...
I think you'd be better-served EQ'ing the HE4XX to a Harman target. HE4XX has low enough distortion to handle a nice bass boost and JDS Atom has plenty of headroom to drive them, even with -7dB preamp for PEQ...