How do Gmasks work in Action?

How do I use 2 gmasks in action to select 2 areas of an image that I want to keep? In the Gmask Tracer you just draw 2 masks, they default to white and everything else is black. In action, drawing 2 masks makes 2 black areas. Inverting both only keeps the intersection. How do I get the equivalent of 2 white masks in action?

Who the hell designed this? :upside_down_face: :rofl:

A colleague just showed me that you set invert on the first mask and then set every subsequent mask after the first to subtract.

So my question is, who the hell designed this?

1 Like

I thought it was just me! :joy:

Definitely invert both masks (obviously, because that’s the opposite of masks in GMTracer :thinking:)
I read once to set them all to Multiply, but same difference I guess.

Because of honouring legacy builds, I guess this will never get straightened out.
Or maybe there is a logic to it which I’m missing.

1 Like

This also seems to work. Everything set to invert and multiply.

It’s so straight forward :sweat_smile: [sarcasm]

I’d love to hear the explanation from Autodesk of why this doesn’t just work exactly like in the Tracer.

I usually invert both and then intersect/ inside. But multiply works… or substracting non-inverted masks…
I do agree it’s not really intuitive. As you said… in tracer, a newly drawn mask is 1 but in action it’s 0… I’ve always found that weird. I do understand where that’s coming from…

1 Like

I guess the problem here is the word ‘mask’

You can add multiple masks in action and they mask the objects.

What you are trying to do is more challenging. You are not masking but trying to add selects.

I’d be happy for them to be renamed if it meant they worked the same as in the Gmask Tracer

or “The Selection Tracer” as it would then need to be called

Gmask tracer is not masking. It is making a matte. White you keep. Black you ignore.

Inside action this is not as easy. Which object are you maksing? Why not modify your matte input.

I get it that it is frustrating that is doesn’t do what the gmask does but the gmask does not have multiple inputs so therefore it is not the same

1 Like

It would be super easy for the masks in action to work exactly like the Tracer:

  • Default with no mask is white (keep everything)
  • Add a mask (selection), and it’s white inside, black outside. (which you can change in the colour field - for some reason absent from action masks).
  • Multiple masks add multiple white or black areas (dependent on the colour field).

Having multiple inputs and layers doesn’t complicate things as you just use Gmask link to choose which images are affected by the mask.

I guess if you just wanted to cut a black hole into an image like the current behaviour then you’d need to be able to select the background colour like on the output of the Tracer node. Or just hit Invert

I agree. Just make a mask outside action and mask the result of your action.

Keeping it inside of action is where it gets complicated because action is so much more advanced than what you are trying to do.

I normally do all my masks outside of action. In fact I normally do most things outside of action. Probably why I’ve only decided to care about this today. :rofl:

However, there are times when being able to mix masks and selectives inside of action is really useful. I just wish it wasn’t so counter intuitive.

I also wish you could copy and paste masks between Action and the Tracer and they would work the same straight away without having to invert everything.

Please Upvote this if you agree:

Gmask Tracer - Consistent reveal/hide property across all Gmask uses

1 Like

I generaly Invert and Multiply in action with a Gmask-Link to the surface when selecting something.
But there’s times when using a gmask to punch through or holding out a particular part for being effected is what one would do in Action.

I think a preference setting in Action would be a good solution.
Say an option for the Gmask to be inverted w/Multiply when creating one using the “Add Gmask Link” menu item(or an assigned shortcut).

I brought attention to ADSK about this weird discrepancy in behavior between tracer and Action gmasks when it was first introduced. The answer was a head scratcher and it was never addressed.

1 Like

Could the solve be that a GMask node in Action just defaults to inverted?

That wouldn’t change the functionality, and would keep existing setups compatible (otherwise we need a preference for ‘old’ vs. ‘new’). As many of us just do go in an invert it, if that just becomes the default setting, it saves us one click but is not an invasive change.

The one downside being that when you do math-ops on multiple nodes, the inversion plays mind games with you in terms of subtract, etc. But that wouldn’t be any worse than it already is.


Yeah no worse then it already is as really I don’t use subtract since it does start the mind games as you mentioned.
I was just thinking a user preference for Action but perhaps you guys are right since probably the default for nearly everyone when using Gmask in action is to invert it?

@GPM I remember when Phillipe did the presentation video explaing the use for gmasks in Action for cutting out or punching through an Action scene seemingly to reveal the background. I remember him showing this in a presentation during a flame user group at Artjail as well in conjunction when demonstrating 3D shapes.

1 Like

I remember that. It was a push to bring more functionality into action. I use it on occasion, but the overall philosophy never took with me. That said, a lot of those functions are very handy on the timeline.

1 Like