I'm making a music video. It's a fairly simple scene currently which consists of 4 collections:
A Kit-bashed 3d scenery collection An actor ('cowboy' collection). This is Prores 444 alpha footage A volumetric animated fog collection A lighting collection When I just render everything on the same view layer, it comes out the way I want/expect.
When I split things off into separate view layers, & recombine them in the compositor, the fog is much less dense, & the image has noise.
I've got three view layers currently, & I'm combining them with alpha-over nodes. I've also tried using colour mix nodes, which give slightly better results, but still don't look the same as the image I get without using the compositor.
My three view layers are:
Background - This is the 3d scene, the lighting, & the Cowboy has 'holdout' selected, the fog is off Midground - The cowboy is on, but the 3d Scene has holdout selected, the lighting is included, the fog is off Foreground - Both cowboy & 3d scene have holdout selected, & the fog & lighting are on. Essentially, I'm trying to re-combine scenery, cowboy & fog in the compositor.
I think I'm doing something (hopefully simple) fundamentally wrong at the compositor stage, or have I messed up the elements in each view layer?
Any ideas? I've shared three images. The 'correct' image which doesn't use the compositor, the 'incorrect' image which does, & my compositor node tree.
Help please, in language a small child would understand! :)


