Currently, a Blend node's output resolution seems to be solely based on the Background input. It would be nice to have an option to either:
Specify which input to pull output resolution from (background, foreground, or mask)
Declare output size as relative to input (current behavior) or absolute (user can set the output size)
Then a scale operation (or some other cheap and fast operation) could be applied to non-resolution matching inputs.
48x48 Background Input
1x1 Background Input
This is mostly a best practice pulled over from my experience with Substance Designer. There it is much cheaper to work with 8x8 solid textures than 1024x1024 (or larger) solid color textures. I'd assume that a similar principal would be true here as well, with a 2x2 node being cheaper memory wise than a 32x32 node. Granted, Pixel Composer sizes are probably much smaller on average and therefore not as taxing so this may not really be much of a performance issue ever.
Even so, supporting more flexible blend output sizes would help speed up workflows since we wouldn't need specifically sized Solid nodes during blend operations.
Thanks for your suggestion. I agree with you on the convenience side since I found myself adding scale and padding node a bit too often in the past as well.
Currently, a
Blend
node's output resolution seems to be solely based on theBackground
input. It would be nice to have an option to either:Then a scale operation (or some other cheap and fast operation) could be applied to non-resolution matching inputs.
48x48 Background Input
1x1 Background Input
This is mostly a best practice pulled over from my experience with Substance Designer. There it is much cheaper to work with
8x8
solid textures than1024x1024
(or larger) solid color textures. I'd assume that a similar principal would be true here as well, with a2x2
node being cheaper memory wise than a32x32
node. Granted, Pixel Composer sizes are probably much smaller on average and therefore not as taxing so this may not really be much of a performance issue ever.Even so, supporting more flexible blend output sizes would help speed up workflows since we wouldn't need specifically sized
Solid
nodes during blend operations.