Export to Null is different than stabilize

Hello,
a basic question: I can of course stabilize my footage.
But instead I wanted to apply these transformations to a null.

But the null has a much bigger (wrong) movement than a correct stabilze “render”.

Attached screenshots. I checked invert and expected the same transformation result.

(the referenz frame is somewhere in the middle, but this shouldn’t have affect on the export of transform data)

I have deleted of course scale and rotate from the Null transform, so both ways only change the position.
Without the null the layer it doesn’t move, so it seems not to have to do double transformations.
Inside MochaPro I tracked only one layer and checked at “stabilze” only X and Y position.

Thanks for help







It makes sense that a null and corner pin stabilize are going to apply differently. Can you tell me what you are going for as the end result effect and maybe I can get you a workaround?

Hello, I have once stabilized footage and instead of applying it directly to the layer, I took the (inverted) transformations and applied them to a null.
So any object, which is parented to the null, becomes stabilzed in the same way.
Now it doesn’t work anymore.
I mean it’s afaik the standard workflow, not just to stabilze footage with the integrated mocha “stabilze - render”, but with applying those transformations to any other object.
I have seen that in your tutorials.
So it makes no sense, that a stabilization is coming out wrong, when I apply it to a null.
Thanks for help

05c Using the Data - Stabilization [Mocha Essentials] - YouTube exacty like here (as far as I remember)

It looks to me like you have tracking data up to skew for the corner pin and x and y only for the transform null, it’s expected that this would be an issue.

If you say “up to skew” this means the settings “shear x and y”? (I’m not very good in English)
I’m only taking the position data, when I delete scale and rotation - actually, it shouldn’t make a difference which information has been tracked.
So you think, if I disable shear, it will work? I didn’t use a corner pin (I tested it but it has been only distorted)

If you track more parameters than translation and then delete parameters up to translation then you’re not going to have a very stable track. The only way to get a 1:1 comparison between your track and the null is to track translation only. It’s unlikely that will lock your shot down though unless it is only unstable in translation (x and y only).

The only way a corner pin will operate the same as a transform is if you’re only tracking translation.

Shear/skew is the same, I used the wrong word.

Okay, thanks, would be great if that’s the solution, I will try it tomorrow!