BIG problem with animation and light cache...please help

Hello guys,

got a big issue here…

rendering animation where a lot of stuff is fading in (animated using graph editor) …various objects…

rendering methid used is IRR map for primary and light cache for secondary.

now…what we get is ridiculous speeds for lightcache calculation where objects are fading in and out…10 HOURS and more !!!

Is this rendering method suitable for this animation type or not?..if not …what is the most suitable method for doing this?..we have loads of moving objects, stuff fading in and out…want clean solution…no flickering

can u please help us as the deadline is approaching and we only found out today :shock:

thanks very much

Martin

Hi Martin,

As far as I know there isn’t a quick way to do fading objects directly out of VRay. It’s far quicker to to render one sequence of frames with the object fully visible and then one qith the object hidden and fade btween them in post. From my experience anyway…

Dan

thanks for good point Dan …that will be a lot faster … :sweat_smile:

the problem still persists in areas where some objects fade in as they move…:frowning:

No, you can still do the same thing… render the segment where the object fades twice. Once with the object on, once with it off, then blend in post… it will render much faster.

Even scanline is slow when it comes to object visibility, and the reason is because the renderer calculates the object twice, and then blends the results based off of the % of visibility.

Now, if anything in your scene besides the camera is animated, you don’t want to be using IRRmap… to avoid flickering you’ll have to set the settings so high that your rendertimes will be unusable.

what do u suggest dynedain?
the camera moves as well :))))

bruteforce?
or

qmc and LC?

but the light remains the same…its is not animated

were rendering the animation to be presented on the dvd…

would u be so kind and suggest any settings that would be most suitable for this kind of animation?

when we have moving objects and moving cameras, we usually use QMC + LC

The rendertimes aren’t great… but they’d be even worse with Irrmap + LC

would i get away with qmc at 24 subdivs? and LC at 1500 …0.2 sample size?

Depends entirely on your scene, resolution, etc…