Blackmagic Forum

MediaIn1 is video footage
MediaIn2 is a PNG

Why is it that adding a ColorCurves1 node after the MediaIn2 also affects the video footage from MediaIn1?



**Prepare for rant**

I have tried learning Fusion on multiple occasions but this thing is a confusing ridiculous platform. The hassles that one must go through to do the simplest of things is ridiculous.

I broke down and made the curved arrow in PShop and brought in a PNG because trying to make a custom shape filled with white with a black stroke was near impossible to figure out. Nowhere could I find any tut or video showing how this is done. In Aftereffects, done in minutes. Fusion, forget about it!

For graphics or text, AE is still king. Fusion (standalone) is great in other things.

Moved to Fusion forum, given its a fusion question

Why is it that adding a ColorCurves1 node after the MediaIn2 also affects the video footage from MediaIn1?

Probably because your CC1 affects your alpha and/or your background color, which - by nature of math - also affects your videofootage. You have to check the pre-divide/post-multiply checkbox in the CC1 tool and you should be fine.

This is not jumping through a hoop, it's a way to allow the tool to work with various kinds of footage.

And maybe it's your algorithm, but when I put this in the search area in Youtube

https://www.youtube.com/results?search_ ... +in+fusion

the top 3 results all explain how this works in videos that are shorter than 10 minutes, so if you say 'nowhere could I fund any tut or video how this is done' then you just didn't search hard enough.

Hi
As you can see from these 3 images, The composite with basic color curve with contrast works fine as in the first image.
watch image on the left screen, it shows alpha transparency as unaffected, you have to keep in mind that when you mangle it out of proportion the alpha transparency is also affected and this is going to color the merge over.
In Image 2 you can clearly see on the left screen that the transparency channel is affected and that's going to color the merge.
Now if you have messed up the alpha transparency, no problems, this is where the Node based compositing really shines.
you can extract the alpha Transparency from the original Image and re-apply it to the merge using Bitmap node.

You have to keep in mind that Node based workflow works a little different than a time based workflow, both have their advantages and disadvantages.
Node base you have more control, but you will come to appreciate the extra control as you try to wiggle out of tricky situations and complicated and complex composites.

Code: Select all

   InCanvas = self:AddInput("Affect Canvas", "AffectCanvas", {
      LINKID_DataType = "Number",
      INPID_InputControl = "CheckboxControl",
      INP_Integer = true,
      INP_Default = 1.0,
      IC_NoReset = true,
      IC_NoLabel = true,
      })   

...

   local do_canvas = ( InCanvas:GetValue(req).Value > 0.5 )
   local canvaspixel = Pixel()

...

   if do_canvas then
         OutImage:Set(req, out)   
      else
         img:GetCanvasColor(canvaspixel)
         out:SetCanvasColor(canvaspixel)
         OutImage:Set(req, out)
      end
   end



Probably makes more sense in common controls, but that's the general idea. Would save a lot of heartache if this was just standard.

Fusion is like a Ferrari. It peels out so much that it's almost impossible to drive. But once you get the hang of it, hold on to your hat.

I still peel out a LOT.