Nor does it means it sounds worse because you have measured more aliasing.
?? We were talking about amp tonestacks and the fact you can't easily replicate their behaviour with a generic EQ, what has aliasing to do with that?
Anyway, sounding better or worse is subjective, some people hear some stuff, some people hear other stuff... But aliasing is measurable hence objective, and it's a thing a real amp doesn't have, so the less a simulation has it, the (objectively) better it is.
My first question was, did you run the software or the pedal. If this problem is solved by more oversampling, certainly then can do that in the software.
I already answered that question. Not sure that can easily be done with AI captures, most likely not with already done ones.
E.g. NAM captures currently can be "oversampled" only by making captures at 96 or 192 kHz, but you can't change that afterwards
Your measurements indicate they are using less than 8x oversampling (because Fractal uses 8x ... they must be using less.)
My measurements indicate they're not using oversampling at all, harmonics reflect right at 22.1 kHz.
And there's other weird stuff caused by a wrongly implemented sample rate conversion too that appears if you run the plugin at 48 kHz.
As I already said, this is all covered in the original thread from last year, so go read that if you want to dive in.