If I’m composting into a low quality shot and using a background video, how would I go about reducing the quality of the objects to match the video without also further reducing the video quality? I did think about rendering with alpha layers and then putting it on top, but that seems inefficient, especially if I do masking and point tracking. Currently I’m just turning the render quality down to 50% overall. Any ideas?
There are many reasons that the background vision may not match your renders. Quality of the video source is just one. Lighting and motion blur as well as render engine and lens matching are others.
If it is just an issue with replicating the look of the video then you need to analyse what makes the video different from the render. This could mean you need to use all sorts of nodes to change the colour of your renders or add compression noise to simulate the artifacts in the video. All of this is quite hard though.
I have determined that it is simply the video quality. I just needed to know what node you would use to add pixelation/compress to make the object fit that lower quality.
Why not render your BG video out with a compressed codec? Say H.264 at 20 mbps as a .mov or whatever you want to match? It may take some matching but that is the art part of it. So try 15 frames and check it until you get it close… Or try some gaussian blur on it Under camera icon when setting up a render out…
Honlad, there is no 1 node to achieve this effect. And when you do affect the media to match the problem will be keying it over the background.
In cinema fx you typically remove artifacts, like noise or grain (in the old days), then add the clean effect shot. Finally you replace the grain or noise over the whole image to make it uniform.