|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
nemesis nous apporta ses lumieres en ce 2008/01/27 20:20:
> so, any free time in your hands? want to put that quad core to stress with
> povray 3.7beta? If you're, try the little scene below. With +AM2 +A0.04 +J1.0
> 1024X768 on a P4 2.66, it just brought my PC to a crawl for the whole weekend!
> It all began with a challenge in a brazilian forum, when some guy claimed this
> scene:
> http://200.213.105.41/greysilica/exemplos/cjcopojateado.png
>
> took him 48 hours to render with Blender internal renderer.
>
> Surely something was wrong, or so I thought. Well, in the region between the
> bottom of second row cups and the top of first row cups, render time was about
> 11 minutes a line! It is all about refractions, normals and fresnel
> reflections. 40 hours total on a P4 2.66 with povray 3.6.1...
>
> I'm sure it's possible to optimize it with some no_image, no_shadow trickery
> with doubles and no_reflection in the main cups, but I didn't thought of it
> beforehand!
>
You have multiple reflections and refractions. That mean that you have tons of
rays to compute. Add the aa and the small bumps, you need lot of subsampling =
lot of time.
Try adding adc_bailout 0.01 in the global_settings.
It will cause a small decrease in quality that can go unnoticed, but will
accelerate things as the ray tracing will often stop much earlier.
--
Alain
-------------------------------------------------
The strongest reason for the people to retain the right to keep and bear
arms is, as a last resort, to protect themselves against tyranny in
government.
Thomas Jefferson
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
seemingly, yes. It gives a "suspicious" warning at the console output.
It's rgbft. I think I should've put some more transmit and a little less
filter. Hmm, I suppose 1 filter and 1 transmit do not work right: the
color became quite bizarre when I tried... I also think the original
image had a lot more diffuse and ambient as well...
Post a reply to this message
Attachments:
Download 'us-ascii' (1 KB)
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> seemingly, yes. It gives a "suspicious" warning at the console output.
> It's rgbft. I think I should've put some more transmit and a little
> less filter. Hmm, I suppose 1 filter and 1 transmit do not work right:
> the color became quite bizarre when I tried... I also think the
> original image had a lot more diffuse and ambient as well...
I heard (filter + transmit) should be <= 1 for things to look correctly.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain escribió:
> You have multiple reflections and refractions. That mean that you have
> tons of rays to compute. Add the aa and the small bumps, you need lot of
> subsampling = lot of time.
> Try adding adc_bailout 0.01 in the global_settings.
> It will cause a small decrease in quality that can go unnoticed, but
> will accelerate things as the ray tracing will often stop much earlier.
>
He also has focal blur (commented out on the scene, but looks like the
image has it).
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> z*2.6-y/3+x*.1 instead of <0.1, -1/3, 2.6>? This isn't the obfuscated
code contest, you know :)
it's what happens when you begin with just one dimension and then add
another, and then, a last one as well... :P
> I'm trying to modify the scene so that the camera gets rotated, instead
> of the whole scene. And taking that lightsource outside the union. My
> final goal is a camera rotate y*360*clock that looks good. It's not
> turning out so easy...
not really needed camera rotation for that.
just before rotate -x*amount
add:
rotate y*360*clock
Post a reply to this message
Attachments:
Download 'us-ascii' (2 KB)
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nicolas Alvarez napsal(a):
>>
>> seemingly, yes. It gives a "suspicious" warning at the console
>> output. It's rgbft. I think I should've put some more transmit and a
>> little less filter. Hmm, I suppose 1 filter and 1 transmit do not
>> work right: the color became quite bizarre when I tried... I also
>> think the original image had a lot more diffuse and ambient as well...
>
> I heard (filter + transmit) should be <= 1 for things to look correctly.
f+t <= 1
f > 0
t > 0
Did I forget something?
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I think he used basically the same setup, but only too late I realized the
light should be closer and be directional. Also, I think he used some
specular on the floor. It's much too bright.
BTW, my bezier curve drawing skills suck. :P
Post a reply to this message
Attachments:
Download 'us-ascii' (1 KB)
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>
> not really needed camera rotation for that.
>
> just before rotate -x*amount
> add:
> rotate y*360*clock
I tweaked lots of settings. I think the camera position doesn't look
exactly as before, but good enough. Rotating the whole scene in the way
you say didn't look right, for some strange reason. I also had to tweak
the z position to make the 'rotation center' better.
[global settings, background, cup, and tex unmodified from original,
snipping to make post shorter]
light_source {
1.6*(4-z*8)+y*8
1.2
area_light
x,y,4,4
circular orient jitter
translate -z*0.8
}
union {
plane { y 0 pigment { rgb 1 } finish { ambient .2 diffuse .6 } }
object { cup translate 0 tex( <.7,.8,1,.8,.1> ) }
object { cup translate z*.7 tex( <.9,.86,.4,.8,.1> ) }
object { cup translate z*(.7+1.3) tex( <1,.96,.84,.8,.1> ) }
object { cup translate x*.7 tex(<1,1,1,.8,.1> ) }
object { cup translate x*.7+z*.7 tex(<.96,1,.67,.8,.1> )}
object { cup translate x*.7+z*.7*2 tex(<.7,1,.7,.8,.1> ) }
object { cup translate -x*.7 tex(<1,.6,.6,.1,.8> ) }
object { cup translate -x*.7+z*(.0+1.3) tex(<.5,.5,.8,.1,.8> ) }
object { cup translate -x*.7*2+z*(-.7+1.3) tex(<.7,1,.94,.1,.8> ) }
translate -z*0.8
}
light_source { <0,1,-8> .7 shadowless }
camera {
location <0,1.5, -3>
look_at 0
angle 55
//yes, I actually tried focal blur for the final render
aperture .04 blur_samples 60 confidence 1 focal_point y/3+z*2
rotate y*360*clock
}
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> [global settings, background, cup, and tex unmodified from original,
> snipping to make post shorter]
>
Well, I did change rgb with rgbft on the texture :)
PS: you're posting in HTML again.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> You have multiple reflections and refractions. That mean that you have
tons of
> rays to compute. Add the aa and the small bumps, you need lot of
subsampling =
> lot of time.
yes, but looking at the original scene for the first time, I'd never have
guessed it!
but you're right, there are rays passing through 3 cups with IOR and
normals, as well
as reflecting off the other cups in the scene. It's insanity squared!
> Try adding adc_bailout 0.01 in the global_settings.
I'll try that. As well as some no_image, no_shadow trickery...
> It will cause a small decrease in quality that can go unnoticed, but
will
> accelerate things as the ray tracing will often stop much earlier.
And I didn't even go with media for blurry shadows, area_lights, radiosity
and caustics!
I wonder how much time it'd take in an unbiased renderer such as Indigo or
Luxrender...
Post a reply to this message
Attachments:
Download 'us-ascii' (2 KB)
|
|
| |
| |
|
|
|
|
| |
|
|