I mentioned my own motivation at the end of my September post. The short version is:
-
In extreme cases, the lack of frame-to-frame consistency made it harder for the stabilization plugin to do its job. (You can see some shaking around 0:09 in this video where it was too much.)
-
Knowing where the sprocket holes are gives an angle that can be used as a hint while calibrating the straightness of the film transport.
-
If you know the sprocket hole location and dimensions, you can infer the frame position and dimensions. This can be used as a safety check to make sure the entire frame is in the camera’s viewport before capturing an exposure. And while you’re at it, you can use the same dimensions to crop out any excessive over-scan (leaving, say, 5% instead of 20%) to save a little disk space and give the stabilizer (and other processes) fewer pixels to chew on.
For my part, the first one is the real reason. The other two are just nice things that you get for free once you complete the first one. 