Anyone using Twixtor?

Trying to test Twixtor against FCPX built-in optical flow. Not seeing Twixtor do any better. Is anyone using it, and have feedback about it?

… years ago, was allowed to 'play' with it for an hour on a friends machine ... very complex interface, seems to allow to 'mask' fore- from background, didn't found out how.-

With FCPX 10.2 (??) OF got an update, some say not, I noticed a dramatic improvement.

Depends on source: avoid motion-blur, try 'simple' background, avoid too high angular speed..

Have shown this example earlier: recorded 1080/100p, no tripod, no solid background. OF with 10% = little wobble at the shoes - far from perfect but usable.

Another example, now all knobs to max: 100p recordig, 1% (one) speed/OF = some wobble at the fast moving parts (hands). Igonre second part, that was shot with a 50€ actioncam ....

Would be nice to see those done with FCPX's Optical Flow for comparison. That's my goal, to see side by side comparisons, see if Twixtor is worth the majorly huge price tag or not. We have shots on our TV show that will be 60fps and 30 fps. Delivery is 60fps. So we can't over-crank in-camera, really. Action shots, lots of movement. Drone shots, helmet cams, handlebar cams, all about snowmobile riders. So trying to test out best way to get our slow motion shots.

I need to run some more tests, but I'm not seeing Twixtor do a better job, just slightly different artifacts.

BenB wrote: Would be nice to see those done with FCPX's Optical Flow for comparison. …

sorry, not at hand, as mentioned, I was invited to spend unaccompanied an hour at some 'pro' system, ... tried to understand the UI (failed), followed the tuts from Twixtors website, results were not overwhelming, but source files were far from perfect (AVCHD, 720/25p), learned later about 'no natural motion blur'. Plus I didn't understandd the purpose of some sliders - never got a better result, my attemps made it allways worse .....-

I assume, snowmobiles = bright sun = mega-crisp source material... :lol:

Only prob is angular speed - if the 'difference' btw. frame#100 and frame#101 is too big, even Twixtor can not do magic.


Most demoes from them are the typically 'fly high'-stuff = solid blue sky as background and some 'healthy' distance to subject to avoid angular speeds...... => I dare to say, then the actual OF can perform similiar.

.... and I never found advice about that 'masking' / background/foreground feature, to tell the software, calculate that, don't calculate that....

Everybody talks about Twixtor, but nobody uses it. Or at last ist willing to share about it....

PS: that tiny production house (3 persons) where I tested it, owned but didn't used it either LOL!

If anyone else reads this, and has a slow motion solution besides these, please let me know.