[andronikos916]

If you let it render more (and you have specified enought time) it will look like the preview at the bottom.

the only thing is if you have added the -mxi:w option or not? if yes you can re-render it without loosing the previous rendering.

..if not then...


cy,
Andronikos

[andronikos916]

set time: 9999999999

sampling level: 50

and you should add after the "-d" the "-mxi:w"

then if you want to rerender it you should have:

-d -mxi:w -mxi:c

cy,
good luck
Andronikos

[andronikos916]

my pleassure Denis.

always happy to help,
Andronikos

[iandavis]

you can't tell from the preview when dealing with emitters... particularly if you have any SSS surfaces, the preview will be very bright and pixelated in those areas. The only reliable way I have been able to get a good render is to render a small image to at least level 10... at that point the output is pretty close to the final. Note, there are certain iterations which seem oddly out of sorts with the preview... I think it's 7 where everything goes quite dark...but the preview doesn't change... then by 11 it's brightening up... by 14 you should have a pretty clear example of the final. Once you get an approximate level then set the final render with that exposure, then, at level 10, compare the preview with the display again... and make adjustments... then again at 14, etc.

So, the lights in your example shoud be visible by 15 but they wont be as bright (most likely) as the in the preview... In other words, use the quick render NOT the preview to get an idea of your emitter levels. I had that problem getting the SSS right on my austin mini tail lights.
http://www.iandavis.ws/maxwell/mini_rear.jpg

Eventually I just ignored the odd glows that showed up in the preview and trusted 2 hour test renders instead... sucks if your on a tight deadline, but after a few times you will get the idea of how the two connect, then you can make better exposure guesses. good luck.


quirks in the beta I'm guessing



btw nice model, I can't wait for you to post the finished product in the gallery

[Renato Lemus]

Andronikos, isn't supposed that this version adjusts automatically the render output according to the rendertime?, in older versions I read that crancking up the time was the right choice because the renderer will stop at the given sampling level.
But in tis beta, I think that if you set time to 5 minutes, the render will look different than if you set 999999 minutes and stop at 5 minutes.

[tom]

No, it's not Renato.

[andronikos916]

if you are using max and type 99999 you will get 72000min. that is normal..

good to go Denis!

cy,
Andronikos

[siliconbauhaus]

best of luck mate.....and hope you dont have an error

[sidenimjay]

obviously the preview pic gets ahead of the actual render, but is there a way we can see what sample level the preview is on?

its very annoying to look at the preview pic and see caustics that dont exist in the beauty render , with no way to estimate how long the beauty render will take to produce the same caustic. i have seen cases where it may take 2 or more days to resovle to the preview, with some notable differences between them.

do the sample levels even correlate ?

at some point as mentioned earlier, the preview gets way ahead of the final render . . .is it 2 samples ahead , 10 , 50 , 100??? and does a sample level of 18 on the preview render look like sample level 18 on the main render? if not, then whats the point of the preview render? anyone with a maxwell license and half a brain can run a test render at a smaller res to test, and it would at least be accurate

does it clock cycles to the render time? or does it filter across to the main render what it does in the preview?

if it adds to rendertimes it would be preferable to have the option of using it or not . . .

[Renato Lemus]

No, it's not Renato.

ok, thanks for the info.

[DELETED]

DELETED

[Renato Lemus]

rennauto?

http://images.google.ch/images?q=rennau ... a=N&tab=wi