Hi, Im trying to figure our how to keep the scale of a camera track consistent in flame.
I can track it well in syntheses, then I set the coordinates, clicking on a set of 3 points to set the ground plane. As I understand it this should set the scale too. So the first tracking point you click becomes the center of the ground plane. The second should set the scale. However when I first did a track and imported to Flame that second axis had an x position of 2000, the next time I did it it had a position of 4000, then 100,000 then 10,000. I’m trying to join two shots and these three tracking points I’m referencing are supposed to set the scale. So I was just wondering if anyone has any idea why the position is always a round number so it is obviously setting the scale by those coordinates, but it is not consistent.
I’m not selecting “Auto fit in scene” when importing the FBX from Syntheyes.
Have a look at your World Size in Syntheyes’ solver window. My guess is that number will be changing scene by scene.
The *3 coordinate picker will set your ground plane and camera orientation, and it will assign a more or less random world scale value, but to my knowledge the only way to make the world scale consistent between shots is to a) set an additional distance constraint by clicking a point then command-clicking a second point and entering a value (repeat with the same points and value for all shots, obv), or b) track all your shots in the same Syntheyes scene with cross-referenced constraints, which is something I’m currently trying (with limited success) to learn how to do.
You might be able to just change your world scale to a common number for all your shots, but I’ve never tried that and I assume it wouldn’t stick if you were to re-solve. Distance constraints were designed for this exact thing.
(Adding…) I’m going from memory, now that I think about it, I am not positive where you will find the world size number, but look for it, because I’m fairly confident that’s where your problem is.
Thanks for that, really helpful. When you say “set an additional distance constraint by clicking a point then command-clicking a second point and entering a value” Where do you enter that value?
Ahh, yes. In the coordinates room, perform that operation, then on the left side there will be a field named “dist” that will have a value of something like 0.0, click that and enter a number (any number really, as long as it’s consistent across your shots).
For further reading/insomnia remediation, here is the relevant page from the the beloved Syntheyes manual:
Like I said, there’s an even more intriguing workflow where one can track a bunch of takes/angles in the same syntheyes setup (track camera 1, go to shot->add shot, disable camera 1, track camera 2, etc) but I haven’t gotten it quite figured out to my satisfaction.
Victor Wolansky’s fxphd syntheyes course has a whole chapter on it, but it hasn’t been enough to penetrate this thick skull of mine.
Is this what you’re talking about Kirk? Linking the known tracking points together across shots? Multi-Shot Tracking
Never used it before, but it seems like an interesting workflow for this kind of thing. The other thing I’ve never dabbled in are the phases. That too intrigued me, kind of a node based system for certain aspects of the track, fov, coordinates, etc.
That’s the stuff. I’ve gotten it close to working, but I think the expected workflow is starting with a survey sequence and then solving indirectly and I just haven’t made it all work yet.
As for phases, I also am intrigued, but it’s pretty clear from the lack of 3rd party tutorials that nobody uses them, which dampens my enthusiasm somewhat.
