|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi all,
playing with (a lot of) photons, here is a recent render with a flint-glass
material, and an isosurface inspired from Mike William's tutorial
(www.econym.demon.co.uk/isotut).
Cheers.
Bruno.
Post a reply to this message
Attachments:
Download 'isophotons.png' (320 KB)
Preview of image 'isophotons.png'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Bruno Cabasson" <bru### [at] cabassoncom> wrote:
> playing with (a lot of) photons, here is a recent render with a flint-glass
> material, and an isosurface inspired from Mike William's tutorial
Wow! How long did that take?
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> "Bruno Cabasson" <bru### [at] cabassoncom> wrote:
> > playing with (a lot of) photons, here is a recent render with a flint-glass
> > material, and an isosurface inspired from Mike William's tutorial
>
> Wow! How long did that take?
I was playing on my small laptop, with a small AMD P920 quad-core @1.6. I took
about 6 hours, 74+ million surface photons were stored. I needed that many in
order to have smooth results in some areas where light is 'diluted'. If you
shoot too few photons, you don't event see caustics in cases like this ...
B.
PS: Thanks for the 'wow'.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Bruno Cabasson" <bru### [at] cabassoncom> wrote:
> "Cousin Ricky" <rickysttATyahooDOTcom> wrote:
> > "Bruno Cabasson" <bru### [at] cabassoncom> wrote:
> > > playing with (a lot of) photons, here is a recent render with a flint-glass
> > > material, and an isosurface inspired from Mike William's tutorial
> >
> > Wow! How long did that take?
>
> I was playing on my small laptop, with a small AMD P920 quad-core @1.6. I took
> about 6 hours, 74+ million surface photons were stored. I needed that many in
> order to have smooth results in some areas where light is 'diluted'. If you
> shoot too few photons, you don't event see caustics in cases like this ...
>
> B.
>
>
> PS: Thanks for the 'wow'.
I take the occasion to say that such scenes would benefit from multi-cores
photon shooting.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 14.12.2012 12:46, schrieb Bruno Cabasson:
> I take the occasion to say that such scenes would benefit from multi-cores
> photon shooting.
Multi-core photon shooting is automatically used /if/ the scene contains
multiple light sources (or maybe multiple photon targets; I'm not sure
about that though). With just one light source and a single photon
target, it would be non-trivial to split up the job.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
clipka <ano### [at] anonymousorg> wrote:
> Am 14.12.2012 12:46, schrieb Bruno Cabasson:
>
> > I take the occasion to say that such scenes would benefit from multi-cores
> > photon shooting.
>
> Multi-core photon shooting is automatically used /if/ the scene contains
> multiple light sources (or maybe multiple photon targets; I'm not sure
> about that though). With just one light source and a single photon
> target, it would be non-trivial to split up the job.
I am aware of this. Thus, I tried to use as many co-located light sources as
cores on the processor. All cores work fully loaded during the shooting phase,
but the subsequent render seems to be slowed down, probably because rays are
tested against all light sources instead on one.
The doc says POV-ray spawns as many threads for photon shooting as cores. I did
not see something about targets. But I can have missed it.
BTW, in extreme photon scenes, the spiral shape of the shooting process appears
on the result, and jitter is not really a perfect solution if you aim at a
smooth result.
B.
Suggestion: would it be possible to spawn multiple threads, and make the normal
photon generator feed them, like a kind of multiplexer?
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> clipka <ano### [at] anonymousorg> wrote:
>> Am 14.12.2012 12:46, schrieb Bruno Cabasson:
>>
>>> I take the occasion to say that such scenes would benefit from multi-cores
>>> photon shooting.
>>
>> Multi-core photon shooting is automatically used /if/ the scene contains
>> multiple light sources (or maybe multiple photon targets; I'm not sure
>> about that though). With just one light source and a single photon
>> target, it would be non-trivial to split up the job.
>
> I am aware of this. Thus, I tried to use as many co-located light sources as
> cores on the processor. All cores work fully loaded during the shooting phase,
> but the subsequent render seems to be slowed down, probably because rays are
> tested against all light sources instead on one.
In this case, yes, you do illumination and shadow tests against each
single light.
Also, the spiraling photons shooting patterns will be right on top of
each others. In such a case, jitter is your only hope of having your
photons hit different locations.
>
> The doc says POV-ray spawns as many threads for photon shooting as cores. I did
> not see something about targets. But I can have missed it.
>
> BTW, in extreme photon scenes, the spiral shape of the shooting process appears
> on the result, and jitter is not really a perfect solution if you aim at a
> smooth result.
>
> B.
>
> Suggestion: would it be possible to spawn multiple threads, and make the normal
> photon generator feed them, like a kind of multiplexer?
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |