Arttu Voutilainen nous illumina en ce 2008-10-27 13:58 -->
> William Tracy wrote:
>> nemesis wrote:
>>> Gee, man!  The render time was already 72 hours for quite low
>>> resolution, let
>>> alone with higher blur_samples.
>> Aren't I a bastard? ;-)
>>
>> I look at the shoreline in the extreme foreground, and I see random blue
>> and green pixels sprinkled around, which bugs me. With higher samples,
>> you'll get more of a smooth gradient.
>>
>> Another option would be to tweak the aperture, reducing the amount of
>> blur, and reducing the number of samples needed to eliminate the grain.
> 
> Yep, I guess I'll do both: lower the aperture and increase samples. Or
> if Alain can explain that blurred transparency to me well enough that I
> can code it, I might try it first :P
> 
>> Then again, some people are fine with the grain. You decide whether the
>> render time is worthwhile. :-)
>>
>> BTW, now that you mention it, what hardware is this running on?
>>
> 
> The 72 hours was on my own computer, AMD 64 3500+. The bigger version
> (now 140:57:56 Rendering line 130 of 720, 89598 rad. samples) is running
> on Eero Ahonen's bit more faster AMD 4450e.
> 
> -- Arttu Voutilainen

Have you ever done any blured reflection?
Now, do the same, but with a transparent object placed close to the camera.
The thing is, you need to keep the number of averaged normals low, as each ray 
will hit the surfaces twice. Averaging 16 normale résult in each ray been 
multiplied by 16x16, or 256.


-- 
Alain
-------------------------------------------------
I bet exercise equipment would be a lot more expensive if we had evolved from 
starfish.

Post a reply to this message