Cause, yes, its directly related to the 2GB known bug, and older MP4Box sources. I'm working directly with Romain from GPAC on this now. If you can upload to rapidshare the source recording then I will see if I can add this to the debugging loop. Solution, I'm not sure as my test recordings have no sync issues of that magnitude after cuts, so your source recording (or a small example) would help.
I expect 50% of the debugging work from this point will be getting audio sync right. It certainly was for 1.1 and 2.0 is whole lot bigger.
More information may also be in the logs.