Question that's been bugging me!

It might be possible to organise… but in an ideal world, you’d take all the photos under similar lighting conditions.
I don’t know if I’ve taken enough photos to try and attempt something… and I’ve only got rather crude photo-editing software.
The closest I can think of to what you suggest is one of the German moth sites where they often show a variety of photos. Have a look at:
Korscheltellus lupulina - LepiWiki (common swift, a macro which is in evolutionary terms a micro)
Opisthograptis luteolata - LepiWiki (brimstone, a Geometrid macro, which seems to be stable in terms of its pattern although there seems to be a rather ‘washed out’ example)
Griposia aprilina - LepiWiki (merveille du jour, a Noctuid macro, which is a bit more variable although easily recognisable)
Diarsia mendica - LepiWiki (ingrailed clay, a Noctuid, possibly the most varied macro - and one which is sometimes quite hard to ID, for me at least).
Mimas tiliae - LepiWiki (lime hawk-moth, possibly the most variable hawk-moth).

One of the problems with looking at ‘collections’ is that dead moths often fade. So comparing live specimens today with Victorian ones in the Natural History Museum might be a bit tricky - although it could be instructive.

Looking at the examples you mention on that German site it seems there is variation (ignoring the fading etc) once you look in detail at the pattern. Initially look very similar but matching up exact bits of black or other pattern they are different.

Yes, that is so.
So I guess the question is: is the degree of variation random or is there still some evolutionary imperative [or other pressure] for the variation to be minimal - especially in the larger species?
I still think there’s a PhD in there for someone!