Technical automation question

Forgive me if this is a “wide” question:

What are the parameters that determine how automation is played back and how are they set?

I’m asking that question because having done two separate tests I’ve found that there is a measurable difference (not necessarily audible) when dealing with automation on the same content. That is to say that there is a sideband artifact when volume automation is applied to for example a sine wave, and in the other test two files, one with rendered (exported/imported) volume automation and one running the exact same automation live, do not null completely - and more importantly the difference varies between playback attempts.

If it helps I can post better descriptions of what I did, but it seems to me that this is all related to how the Nuendo audio engine deals with automation over time.