|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Here are some considerations about photons I'd like to share.
The left image is the normal render you get with one light source. In
my
opinion, the dark shadows cast by the glass are to dark. I have
observed
this kind of refractive caustics in some experiments I have done with
real
glasses and lights, so they are right from the physical point of view as
remarked by N. Kopp elsewhere in this newsgroups. In the other hand,
the
reflective caustics are to bright.
One extreme solution to both is to shut down photons all together. We
all
know that in most scenes this is a bad solution. There is an
intermediate
solution. Turn on a second light source (in the same position off the
original one) with reflection and refraction off. The result is shown
in
the right image. The light that shoots photons has rgb .12 and the new
one, rgb .88 (this settings may not be the best ones). The rendering
time
of the two images are more ore less the same. The slowing down due to
the
two lights is partially compensated by less sampling needed in anti-aliasing.
BTW, here are the settings for the glass:
pigment{rgbf <.94, .94, .94, 1>}
finish {ambient 0 diffuse .5 reflection 0.065 specular 1 roughness .005}
interior{ior 1.5 media{absorption <1, .75, 1> density{rgb .01}}}
photons{target 1 reflection on refraction on ignore_photons}
They work well with global_settings{adc_bailout .05 max_trace_level 100}.
Finally, this are only my opinions and you may have other ideas/solutions.
Alberto.
Post a reply to this message
Attachments:
Download 'ephot.jpg' (32 KB)
Preview of image 'ephot.jpg'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hrm. While I think the shadow/caustics look better on the right side, I like
the glass itself better on the left pic.
--
Paul Vanukoff
"Alberto" <jac### [at] usbve> wrote in message news:39DF469B.BA9559C2@usb.ve...
> Here are some considerations about photons I'd like to share.
>
> The left image is the normal render you get with one light source. In
> my
> opinion, the dark shadows cast by the glass are to dark. I have
> observed
> this kind of refractive caustics in some experiments I have done with
> real
> glasses and lights, so they are right from the physical point of view as
> remarked by N. Kopp elsewhere in this newsgroups. In the other hand,
> the
> reflective caustics are to bright.
>
> One extreme solution to both is to shut down photons all together. We
> all
> know that in most scenes this is a bad solution. There is an
> intermediate
> solution. Turn on a second light source (in the same position off the
> original one) with reflection and refraction off. The result is shown
> in
> the right image. The light that shoots photons has rgb .12 and the new
> one, rgb .88 (this settings may not be the best ones). The rendering
> time
> of the two images are more ore less the same. The slowing down due to
> the
> two lights is partially compensated by less sampling needed in
anti-aliasing.
>
> BTW, here are the settings for the glass:
>
> pigment{rgbf <.94, .94, .94, 1>}
> finish {ambient 0 diffuse .5 reflection 0.065 specular 1 roughness
.005}
> interior{ior 1.5 media{absorption <1, .75, 1> density{rgb .01}}}
> photons{target 1 reflection on refraction on ignore_photons}
>
> They work well with global_settings{adc_bailout .05 max_trace_level
100}.
>
> Finally, this are only my opinions and you may have other
ideas/solutions.
>
> Alberto.
----------------------------------------------------------------------------
----
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Alberto" <jac### [at] usbve> wrote...
> Here are some considerations about photons I'd like to share.
>
> The left image is the normal render you get with one light source. In
> my
> opinion, the dark shadows cast by the glass are to dark. I have
> observed
> this kind of refractive caustics in some experiments I have done with
> real
> glasses and lights, so they are right from the physical point of view as
> remarked by N. Kopp elsewhere in this newsgroups. In the other hand,
> the
> reflective caustics are to bright.
The problem is this:
Ray-tracing assumes (incorrectly) that all of the light in the scene comes
from a single (or multiple) point light sources. In reality, much
(sometimes the majority) of light comes from indirect diffuse interaction.
To compensate, ray-tracing artists usuall turn up the intensity of the light
source so that it is brighter than it should be. This compensates for the
lack of diffuse interaction. Of course, with the light source being too
bright, the caustics will be too bright. Also, with all of the light coming
from one point, the shadows are completely void of light. In reality, the
light source would be dimmer (making the caustics dimmer), and diffuse
interaction would make everything (including the shadow areas) brighter.
Your solution is good, though it would be better if your "fill light" was a
shadowless light.
-Nathan
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <39DF469B.BA9559C2@usb.ve>, jac### [at] usbve wrote:
> The left image is the normal render you get with one light source.
> In my opinion, the dark shadows cast by the glass are to dark. I
> have observed this kind of refractive caustics in some experiments
> I have done with real glasses and lights, so they are right from
> the physical point of view as remarked by N. Kopp elsewhere in this
> newsgroups. In the other hand, the reflective caustics are to
> bright.
Refraction affects *all* the light that passes through...there isn't any
light that passes straight through a lens without being bent. The areas
that have light refracted away will be completely dark, this is
physically correct.
However, in the real world, you usually don't have all the light coming
from a single source, but have light bouncing off of the surrounding
objects, walls, and coming from other light sources. Also, complex
shapes can refract light into areas that are mostly dark. Try putting
white walls around, and turning radiosity on. A dim light source at the
camera's position would help too, and you might try adding small
irregularities in the glass normal.
--
Christopher James Huff
Personal: chr### [at] maccom, http://homepage.mac.com/chrishuff/
TAG: chr### [at] tagpovrayorg, http://tag.povray.org/
<><
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nathan Kopp wrote:
>
> Ray-tracing assumes (incorrectly) that all of the light in the scene comes
> from a single (or multiple) point light sources. In reality, much
> (sometimes the majority) of light comes from indirect diffuse interaction.
Turn on radiosity, use realistic texture and light source parameters
and the "problem" magically disappears.
______________________________________________________________________
Kari Kivisalo http://www.kivisalo.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Have you tried it with an area light? Might make the shadows look better.
-Mike
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Paul Vanukoff wrote:
>
> Hrm. While I think the shadow/caustics look better on the right side, I like
> the glass itself better on the left pic.
>
> --
> Paul Vanukoff
>
May be because the left one seems to have more contrast?
Alberto
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
MikeH wrote:
>
> Have you tried it with an area light? Might make the shadows look better.
>
> -Mike
Yes I have. To my surprise area lights didn't slow down the rendering
to much, but I didn't like the results.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Chris Huff wrote:
> ... Try putting white walls around, ...
They already are.
> ...and turning radiosity on. ...
I'm working on it :-)
> ... and you might try adding small irregularities in the glass normal.
I have tried it in the past, but I can't stand destroying the beauty of
a perfect cylindrical surface -(
Alberto.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Nathan Kopp wrote:
> The problem is this:
>
> Ray-tracing assumes (incorrectly) that all of the light in the scene comes
> from a single (or multiple) point light sources. In reality, much
> (sometimes the majority) of light comes from indirect diffuse interaction.
>
> To compensate ...
> ...In reality, the
> light source would be dimmer (making the caustics dimmer), and diffuse
> interaction would make everything (including the shadow areas) brighter.
>
Interesting. Your remarks remained me not to overlook the limitations of
point light sources.
>...
> Your solution is good, though it would be better if your "fill light" was a
> shadowless light.
>
Well I was focussed in a slow time rendering solution. Shadowless light
is a faster one.
I have start some experiments with radiosity. Its amazing how radiostiy
can change an image. With some settings you can make the glass glow.
Alberto
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |