Maybe I'm doing something wrong, but I was expecting this to work...
convert tool on vid1.mp4, vid1_alignments.fsa works fine
cp vid1.mp4 vid2.mp4; cp vid1_alignments.fsa vid2_alignments.fsa
convert tool on vid2.mp4, vid2_alignments.fsa skips all frames???
I was experimenting with the difference between editing masks for obstructions and NOT editing masks for obstructions, with a short destination video.
I started by extracting an alignments file from the video (call it vid1.mp4).
I then ran the manual tool, and fixed the landmarks (ignoring the obstructions).
I then copied vid1.mp4 and vid1_alignments.fsa to vid2.mp4 and vid2_alignments.fsa, to avoid having to repeat the previous work.
I then opened vid2[_alignments] in the manual tool, and did the manual mask edit for the obstruction (the landmarks are already corrected, from my previous efforts).
While everything looks good in vid2 (all frames are there and have alignments), when I run the convert tool, it skips all the frames and doesn't apply the conversion.
Just in case I did something wrong and clobbered the alignments, I did a clean experiment where I JUST copied video/alignment to a different name (vid3 instead of vid1 in the same path), and it still didn't work.
Is there some kind of "source filename" encoded in the alignments file, that makes it so when I make a copy of the alignment/video to a different name, it fails to find stuff? Like, maybe it doesn't just encode "frame 1 alignments: <foo>", but it's more like, "frame 1 alignments for file vid1.mp4"... but that doesn't really make sense. Plus, I extract in the cloud, and was doing convert on my local machine, so the absolute path of the source files during extract is different anyway (though relative path is the same I suppose).