There is a front/matte offset MX which is very handy. I’m curious if there could be an offset MX that is driven by a strength alpha/gmask input?
So if the offset is say -100 on X and scale of 200, the pixels in white make the full transform, black no transform, and gray in between the values. All the axis parameters like shear, rotate, etc. would be applicable as well
Kind of like displacement I guess, but more precise and mathematical.
Not sure if this is doing exactly what you want, but I think you can more or less achieve what you’re looking for with UVs. If you take a UV into action, front/matte offset, then pull out an action gmask to drive the strength, and apply the UV to your image downstream it should apply the offset with strength being driven by the softness of the gmask.