That's a lot of THD noise, but again, that was most likely done on purpose.
Yes, Tyll confirmed this in a little audio debate, sorry I'm too lazy to source it.
And yup that really explains the laid-back midrange, and especially the lower mids; that's a pretty deep dip. Most of the THD noise is also in this region.
The upper-mids are slightly emphasised relative to the lower-mids, as I have observed (around the 1-2 kHz region).
I can't explain what causes the lower-treble to sound sibilant though. Is it that peak at the ~4 kHz mark in the raw data?
The laid-back upper-treble might also be due to the steep treble slope on the rightmost portion of the graph. A high amount of THD noise is found in this region too.
The channel imbalances in the upper frequencies is also confirmed for me (it wasn't an issue though).
I can only guess but maybe it is the rising distortion after 5kHz.
The shape of the 300 Hz wave definitely matches with sibilant-sounds, or edgey treble.
Sibilant might be a stretch, The initial spike relating to treble is slightly lower than that of the Amperiors, for example.
I'll call it, a solid treble presentation with possibly concerning high frequency distortion, but its hard to tell from that plot .