Besides the question of sample variation which could play a role in your context, it isn't unheard of that headphones show a number of features at lower frequencies on ear simulators that don't occur to quite the same extent on real humans.
Ex : the ripples you see in the K550 measurements below 200Hz on ear simulators, vs. the rather different results on real humans in that article (even for the ones who got a good seal) :
https://www.grasacoustics.com/files...mprovedMeasurementofLeakageEffects_Harman.pdf
Or the "pad bounce" effect here :
https://diyaudioheaven.wordpress.com/headphones/earpads/
- which you'll never observe to that extent on Rtings' real humans measurements for the same models.
As a starting point a more likely behaviour, as I suggested, could be related to the pads warming up and the headphones behaving in a non-linear way under pad compression (for reasons I'm not qualified to explain).
Some examples here :
https://audiosciencereview.com/foru...er-hd560s-review-headphone.29603/post-1046142
A longer post that illustrates why, in my opinion, evaluating headphones' behaviour under pad compression, and not just breach of seal, can provide quite useful information :
Forgot to give credit to @Sean Olive in his tweet here. I believe this is an actual production sample - it slightly differs how it measures from the production validation sample posted earlier. What @Sean Olive thinks about Brent Butterworth´s measurements at Soundstage Solo...
audiosciencereview.com
I'd love to measure with in-ear mics DCA headphones to be quite honest. They seem quite a bit different from other planars.