Loudness is a bit problematic because many things play a role here.
It is the frequencies itself as you see in the curves of equal loudness (Fletcher Munson )
and as you see in these curves they change with loudness.
This means, even if you have a normalizing algorithm that takes these curves into account they would have to be dynamic to fit each
loudness. So the poweramp and cab would have to be taken into account also.
That alone would be pure horror to programm AND would work only with a setup where every part is measured well.
As a summary, normalizing with a signal generator and only peak values will not work that well because of that .
The other problem is LUFS. Measuring a peak signal like on a uncompressed d.i. signal will not give you the same loundness
even if the compressed and distored tone would have the same peak values.
So:
The solution would be a a algorithm that sends a linear sweep and some impulses through your rig und measures the LUFS on the other side.
The measuring would have to take into account at least a bit the fletcher munson curve. Like the A-Weighted curve ...
But even this would not work well at any volume. Kemper would have to make a Microphone to make it really work at any venue etc ...
I fully understand you and i had the same feature request years ago but in the meantime i don't think it would work well.
here are some explanations https://de.wikipedia.org/wiki/Frequenzbewertung