Best practice and workflow using Smart Tempo on audio recordings with wildly variable tempo

[paradox148]

In a recent thread, we were discussing Beatmapping on audio recordings with many tempo and time signature changes, Atlas007 had suggested to me that:

 

"When using Smart Tempo, better results are obtained when proceeding with regions, one at the time, instead of a long stretch.In the same mindset, also dividing the long stretch into regions according to the (drastic) tempo's and time signature's changes is important to obtain more reliable Smart Tempo analysis and results."

As I tend to work a lot with adding MIDI tracks to existing recordings(many complex structures with lots of tempo/signature changes), know how to do this efficiently in Smart Tempo rather than the fairly tedious beat-mapping would very helpful.  I am creating a new topic as the other one was more about Beat-Mapping specifically.

I started to experimenting with doing this but getting very frustrating results.  As the original file is often one long one, and even if I cut the region into multiple ones, when analysing/editing(SmartTempo) one region affects all the others.  Or: if I start to make the region follow the project, this tends to cause the individual pieces to leave gaps between them(or overlapping).

What would then be the best practice and workflow when doing this?  (cutting long region into multiple ones, then bounce each one to be analysed and edited, then realign and join them up again?  If so, wouldn't using the good ol' Beatmapping be faster?)