Motion Vectors Still an Issue in RS
-
Currently running C4D / RS 2026.0.0 on both my M-Series Macs.
I have the worst time (extremely slow frame renders) when I try and add motion vectors to my final animation renders.
On a whim, I just tested a project file, by rendering a motion vector multi-pass from the Standard Renderer. It looks a hair different than that, from RS, but still seems to work with RSMB (the whole point of doing so) and took a fraction of the time to render.I will clean up my text project, and post here shortly. But I just don't get it. Anyone else have major slowdown issues trying to generate MV's?
Edit: My apologies, I was thinking that I was composing this question over on the Redshift help forum. I'm going to do just that now, but I'll also leave this here.
-
Hi entry-newspaper,
I haven't checked in the RS forum, but switching off the Global Illumination and setting the Sampling to 1.0 cut a simple test scene in half, timewise. Another run, even with all lights off, even the Default Light made it to 1/3rd.
I'm sure you get more tips in the RS forum.
Cheers
-
Thank you, Dr. Sassi. Yes, indeed, I've previous experimented with eliminating nearly every non-essential setting, when trying to output the RS MV pass. Unfortunately, at least in my experiences then, it still took quite a long time - particularly, when working with C4D Particles. And therein lies the issue. If I already have a long primary "beauty" render, I'd sometimes rather just include the MV pass in that, as opposed to a second standalone MV render, even if that is a little quicker.
But I wasn't expecting the good ol' Standard Render MV pass to be so lighting quick. Again, it's not the same looking as that from Redshift, but it still seemed to convey some good data to RSMB.
I'll keep experimenting, as always.
-
Thank you very much, entry-newspaper.
Yes, the time difference is there, and it is a more than "noticeable" one. Of course, the engines are different, and how they apply, for example, displacement.
I'm not sure if I can find more than reducing the time to a 1/3rd, but of course, that is always project dependent. Perhaps exploring increasing the Samples to a very high number might help.
I will check later what the RS forum shared.
Cheers
