1. are you sure the starting timecodes of your assets are being modified? Changing the startign timecode of the sequence in FCP will not necessarily carry all the way through to the mpeg encoding process.
2. might there be a mismatch between the timecode formats of your SCC data and the assets? If your asset is getting encoded as drop-frame, and the original SCC data was generated for a non-drop master, your asset will not have certain frames, for instance 01:01:00;00, that might be in the captioning data file.
This can also result in unexpected offsets and drift between the caption and the program, which often increase as the starting timecode of the asset gets higher.
This might be hard to check, since the SCC file will be in hexadecimal, which, I don't know about you, but I find extrememly diffculy to read. But if the master you captured from using FCP was identical to the one used for the captioning, you should be able to check whether the clips are drop or non-drop in FCP. If you find you have put ndf clips in a df timeline, you'll have to re-encode, but it should work better.
Hope this helps -
Max Average
