Camera Analysis Color Space

Does the (relatively) new Camera Analysis prefer a certain Color Space? It should be agnostic since it’s using Motion Vector data?