Hi Tim,
Thanks for the file - I am definitely able to hear the problem.
The general case seems to be as you surmised - the tempo changes (both slower and faster) in some cases are applied too far ahead. The caveat that I'm seeing is that it occurs only after some length of measures (which I haven't ferreted out yet). There is a tempo change at measure 32 which does not cause earlier measures to suddenly change tempo, but later on in the file tempo changes are implemented ahead of the graph/event of the tempo change. For example, there is a tempo change at measure 225. If you play back the file starting around measure 210, then that tempo change actually kicks in at measure 214 rather than at measure 225 where it's supposed to. The same problem occurs as you mentioned in your post.
The problem has probably been around for a while, but since very few users create such long files, it has just not shown up. I haven't been able to come up with a good work around for you at present, other than to manually go in, ferret out where the incorrect tempo changes are occurring, and adjust accordingly to fit the music better. It's a totally unacceptable (from a programming point of view) workaround, but at present I have no better one to offer. I'll write this problem up for fixing, but we're in transition right now (looking for a developer since the
passing of Mark Walsen) so I can't make any promises about when the problem will get fixed.
Sorry for the bother here. This sounds like a great choral medley to perform - I wish I were singing with you folks!
ttfn,
Sherry
task 2426