But as soon as I turn on my matte clip, instead of just exclude my track mask (as it should normally do with another track mask just above it), It just put the track mask as the invert of my matte clip and doesnt even care about the track mask…
So, the new update just released yesterday and as we saw in the Release Notes :
MO-12058 Tracking a layer shape with a Matte Clip does not exclude any mattes for layers above
So we were happy to test that.
And surprisingly, it works… but not 100% correctly.
Ok, if we IMPORT our roto as a new clip and use the correct clip frame offset, we have our roto moving and then we can use it as a matte clip to exclude others layers. Great !
BUT, if we use the INSERT LAYER instead, the roto is not animated at all. In fact, I think the clip frame offset might be the problem because I have only the first frame of my roto (and the clip starts at 1001)
PLUS, the INSERT LAYER method doesnt show any track matte (using IMPORT method it works)
this is showed in this gif below (click on it to play it)
But the import method is more painful than the insert layer because it forces you to export your roto into a new render file.
Can you please clarify why would you want to use insert layer as a matte clip? What is your goal at the end?
Matte clips are expected to be used with rendered clips/sequences. If you just want to exclude some region from tracking area, you don’t need to render/reimport it back to project. You just need to place the exclusion layer on top
Using the insert input like this is neat, but I think we could improve the workflow by adding some Nuke-specific functionality into Mocha that exposed the different layers from the input clip as different clips in Mocha. Then you could place your roto in the mask layer (or any custom layer) and Mocha would make it available for any purpose. It looks possible to do technically, using a similar interface to the one we use for stereo support in Nuke.