Actually I did this. I had a song recorded all on keys on VST 5 (my machine which my friend eventually took over). Then I moved to SX3 and wanted to re-record it. So not exact, but many of the same sounds, layers, etc…
We both used 32Float/48KHz at that point, and we both noticed a difference in the fullness of the sound (I won’t say the purity because they both sounded good). My friend chalked it up to “better resolution”, which in music is one of those unexplainable terms which doesn’t really help prove anything. I wish I still had the old version to post as a comparison. That would have helped a lot.
Now, realistically, music recorded at the same depth and rate should be exactly the same as any other music at that depth and rate. I always wondered after the fact if VST 5 had a imperfect internal clock, which would miss the occasional sample. I do know that the old Pace/iLok dongles were engineered in such a way back then, that a lot of data had to be pumped through them continuously which really affected performance. The old catch phrase was “buy a legit version, but use a cracked version”, because cracked versions didn’t require the Pace software.
Hehe… those were the dayz.