Transform on zoomed portait 4:3 image results in jumpy / jittery / unsmooth motion

Help
21

Thanks a lot for your investigations. I will definitely have a look at this and reply in the bug report as soon as possible!

1 Like
22

Should I copy the results in this discussion to the bug report? Happy to do so if that’s desired.

23

You could post a link to this thread there, but I wouldn’t bother copying it all over. It’s been seen by everyone who needs to now. Thanks for not letting me handwave it away and doing the legwork on this one!

Thanks @berndmj - I thought I’d seen your fingers in the appimage creation pie at some point … not sure what I’m remembering there then :slight_smile:

For (even if only my!) future reference, this is the repo I was thinking of:

craft-blueprints-kde/kde/kdemultimedia/kdenlive/kdenlive.py at master ¡ KDE/craft-blueprints-kde ¡ GitHub

And that recipe for building kdenlive includes mlt as an ‘external’ lib, with its recipe here:

craft-blueprints-kde/libs/mlt/mlt.py at master ¡ KDE/craft-blueprints-kde ¡ GitHub

If you look at line 21 (in the version current today), that shows us the commit hash of the version of mlt to pull from git for the build:

 "https://github.com/mltframework/mlt.git||c8961b4348d7813ec5626841fe913435b092c17b"

Which is where this stops being entirely trivial and deterministic for our use here, because there’s no historical tags in the craft blueprints repo, and it doesn’t look like it’s been designed for anyone to reproduce a specific historical build, the state of its HEAD is just what will happen if a build is run today …

But if I snoop through the changes commited for mlt assuming the datestamps on the appimages I have are correct:

  • 22.08.1 was created 10 Sep 2022, using mlt 87347147eac8adf5ecfac0c3c4ec2d4aaf041f9d (which had been the version in use since 4th July)

  • 22.08.2 was created 18 Oct '22, using mlt 5e3a0db75b40e2174487938ae08d0cdc6dd50ab3 (after bumping it a couple of times in the interim to pull in new changes to mlt)

And they bracket the commits I threw my suspicion at in the post above.

What’s still not clear to me yet is exactly what problem JBM was originally aiming to fix in them (so I hope he remembers and sees how to fix this without accidentally reintroducing that one! :smiley: ) - but I believe that problem was Real too, so we can’t just blindly revert those changes.