Maybe is your footage on linear? its well know that linear and trackers dont love each other… if so do a CM to rec709 and do your track and then work the rest on linear
Not sure if that’s your issue, but I’ve done a lot of manual tracking lately, and have observed something similar.
You set the tracker and hit F13 to advance the tracker. It’s supposed identify the point and then update the track and shift keyframes. But somehow it doesn’t do that until you hit F13 a second time. The tracker always seems to be lagging by one frame, and record the keyframes only after you’ve advanced one frame further.
Definitely a quirky tracker UI. There have been case where I’ve gone back afterwards and manually moved a point back in place that seemed weird.
Generally not a bad practice to do some manual tracking to understand how the tracker works more thoroughly.
drives me mad too, as its just plain annoying. I do the disable tracker 2 and do each separately. Sometimes reframing the source area of the tracker makes it stick better rather than leaving it default
I never, ever track both trackers at the same time and have never had this problem.
Maybe a personal throwback to the 90’s when track analysis could take 30 mins plus.
In the tracker UI there’s a ‘Image Type’ dropdown that defaults to ‘Auto’, but the other options are ‘Other’ and ‘Scene Linear’. Does that make a difference in this case, rather than doing manual CM?
Maybe ADSK fixed the underlying issue, but ‘Auto’ doesn’t always pick right?
I don’t think so, at least based on my own work for the past month. I’ve been tracking a lot of linear footage and all of it tracks much better when converted to rec709.
Looks like there’s already an improvement request: FL-02836
I’ve added my vote anyway, but it feels to me like this is a bug. If it was just the linear space causing problems, then surely T1 solo would fail as often as T2 (which it definitely doesn’t), and T2 would still struggle to track even with T1 disabled (which it doesn’t so much)?
Don’t want to hijack the thread, but more generally the stabilisers really need some work (and to be merged into one). There are already loads of FRs covering different aspects of it, which may dilute the focus… Here’s one which already has 26 votes: Fl-02382
Yes, that seem like a achilles heel of the current FR system. Often there are several ones for the same core feature. And with the votes splintered between them, I’m not sure any of them rise to the surface.
Maybe there could be a follow-up to the Reverse User Group from a few months ago, but instead of new ideas, we’ll make that a FR triage perspective to help put perspective on existing ones and their priorities from the artist view.