Animation is more a bucked sampler kind of job imo. More control over the max samples a pixel would possibly get. Its also faster most of the times.
Except, you want to use the resumeable rendering workflow, but for that you’d need alot of dive space.
I don’t really know how to explain. The progressive one tents to put hundreds or even thousands of samples on small highlights or light borders. While the bucket would reach its cap of 25 or what you put in the max subdiv spinner and go to the next pixel.
The progressive sampler tries to reach different noise thresholds for the whole image through the rendering process.
Therefor it needs to be able to put a way higher amount of samples in total on the image.
You are forced to get that highlight to a noise threshold of 0.01 even when you’d be more than happy with some pixels at 0.5 because, its white anyway and you don’t care if it has a float value of 10 or 11.
I hope you understand what I’m trying to explain.
I almost always use the bucket sampler for the final rendering when I do animations. I compared the render times a few times and were always able to render at least a little bit faster with the bucked sampler than with the progressive one. Not too drastic in most cases but every second counts if you got to get thousands of images rendered and a hand full of seconds on a strong workstation can be minutes on older render nodes.
Thanks ! Yes !! I clearly understand.
I have to say (I’m a bit lazy since I use with “universal setting”) and let progressive do the job. It’s ok for stills but .. yeah, every second counts for animation !
I’ll try with bucket sampler we will see…
Thanks again