Why IrMap over Brute force for animation rendering?

[coolhand78]

even with the LC Scale to"screen"??

and moreover, is there a way of pre-calculating the lightcache for all the frames, saving it out and then loading it as was the case?
reason being it seems laborious to calculate light cache and go through that whole process if it could be done prior for all frames and then loaded for rendering?

[^Lele^]

even with the LC Scale to"screen"??

That is how it should be,yes, and how it is as by defaults.

and moreover, is there a way of pre-calculating the lightcache for all the frames, saving it out and then loading it as was the case?
reason being it seems laborious to calculate light cache and go through that whole process if it could be done prior for all frames and then loaded for rendering?

Not one which would involve some scripting, and a double render pass, with a pre-render script.
The whole point is to forget to precalc a thing, and there is no saving worth of notice in pre-calculating versus per frame, if the total samples are the same, while the number of constraints for the pre-calculated method is higher: ie. you won't be able to move that vase visible only at frame interval 345-365 without a FULL recalc, any small error in any frame will force another full recalculation from the very start, and so on.

[tricky]

This is a revelation to me as a long time user of VRay - no more pre-calc??? And I can even render a door opening or a light moving??? Get out of town! I've got to give this a try!

[Vizioen]

Just rendered a 2 minute animation with this. Quite fast and complete artifact- and flicker free.

[^Lele^]

This is a revelation to me as a long time user of VRay - no more pre-calc??? And I can even render a door opening or a light moving??? Get out of town! I've got to give this a try!

This attitude of curiosity alone, regardless of the final results, has me in tears of joy each time. Thanks!

Just rendered a 2 minute animation with this. Quite fast and complete artifact- and flicker free.

While this had me go .

In regard to speed: remember that what V-Ray does with this approach can be considered around about the minimum effort towards an artifact-free render.
If you can live with artifacts, you can likely shave time with all the other interpolated approaches.

[AlexP]

Why on earth is that old method STILL on the help pages

[Companioncube]

so just so i understand. no matter the type of animation, moving camera, moving objects or both i just leave on default GI settings and just increase lightcache samples depending on the length of the sequence?

[^Lele^]

Why on earth is that old method STILL on the help pages

A case of development overtaking documentation. :P
Working on it.

so just so i understand. no matter the type of animation, moving camera, moving objects or both i just leave on default GI settings and just increase lightcache samples

Yes. 3000 is generally more than enough, although some corner cases may require more.

depending on the length of the sequence?

Nope, just 3000. Nothing of each frame knows it's inside a sequence, as each frame calculates alone.
Because each is a very exact solution, they can live happily by each other during playback without ever having met beforehand, so to speak.

[jasonhuang1115]

Nope, just 3000. Nothing of each frame knows it's inside a sequence, as each frame calculates alone.
Because each is a very exact solution, they can live happily by each other during playback without ever having met beforehand, so to speak.

Glad I found this thread...
@Lele, will the LC subdivs affect render time? If yes, is it worth to set the value to 1000 or lower for preview render and switch back to 3000 for final render?

[^Lele^]

Glad I found this thread...
@Lele, will the LC subdivs affect render time? If yes, is it worth to set the value to 1000 or lower for preview render and switch back to 3000 for final render?

Sure.
If you can live with artefacts and lower quality, you can both lower the subdivs count, and even turn off retrace entirely for very quick (but incorrect) renders.
In general, however, the standard still-frame defaults will do a decent job even on animations, without horrible artefacting (perhaps some bubbling, and a odd glint here and there on some frame.).
Those, coupled with a higher Noise Threshold (0.1++) will get you quite quick, yet well readable, results most times.

[jasonhuang1115]

Thanks for the tip Lele!

[kosso_olli]

In some cases the Light Cache gets stuck here on our end. Have that problem right now. I have to set the bounces very low, to something like 10 to render. Otherwise the LC never finishes. Well, after 20min I canceled.

[^Lele^]

do send a scene.

[kosso_olli]

Will do tomorrow. Directly to you? Do you have your own account at chaos?

[^Lele^]

I do, name.surname @ the company domain.
Sending it to tech support, however, is just as good, if not often much better.
I can track it anyways, but it's safer in terms of the issue not getting lost should i get accidentally thumped on the head and wake up smart, and should the issue be more than scene setup, they surely have more skills, tools and access than I to get it sorted.