 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
OK, well, I've never tried to do a scene using radiosity before, so here
goes...
As I understand it, anything with a non-zero ambient setting will in effect
"glow". My experiments seem to back this up. Only trouble is, they don't
glow very much; I'm having to use things like "ambient 10" to get enough
light in my scene! (That's fine if the thing is white - doesn't work too
well for pastel shades though.)
Also, the images are very patchy. Now, when I just dunked an empty
radiosity{} block into my scene, I wasn't suprised that the defaults didn't
magically work perfectly. But... I've been fiddling with the settings for
ages, and it's still blotchey. I've managed to improve it quite a bit, but
it's still there. POV-Ray is obviously trying its hardest to smudge out
those pixel artifacts from the pretrace step... but now quite suceeding.
Oh, and about that pretrace step... It seems to interact with the normal
mosaic preview INI setting... Is this correct? Does changing the INI
settings affect the final image, or only the preview process?
Any help would be appreciated. I can post an image if it helps - but
basically it's a large empty room with a smallish opening in the roof that
lets in the light from a superambient object outside. My radiosity settings
at the moment are:
radiosity
{
adc_bailout 0.01 / Cheat
recursion_limit 5
count 400
nearest_count 7
error_bound 1.5
pretrace_start pow(2, -6)
pretrace_end pow(2, -9)
}
(It's a giggahertz CPU ;-)
Thanks.
Andrew.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Andrew Coppin wrote:
>
> [...]
>
> Any help would be appreciated. I can post an image if it helps - but
> basically it's a large empty room with a smallish opening in the roof that
> lets in the light from a superambient object outside.
> [...]
This is never a good idea, you will need an extremely high count (probably
larger than the maximum of 1600) to get smooth results in such a
situation. A high quality area light instead of the ambient object is
usually a better solution.
Christoph
--
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 15 Nov. 2002 _____./\/^>_*_<^\/\.______
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Andrew Coppin" <orp### [at] btinternet com> wrote in message
news:3dd56a62@news.povray.org...
>
> As I understand it, anything with a non-zero ambient setting will in
effect
> "glow".
Well, diffuse value is key now in POV-Ray 3.5. Was ambient before.
> Also, the images are very patchy. Now, when I just dunked an empty
> radiosity{} block into my scene, I wasn't suprised that the defaults
didn't
> magically work perfectly. But... I've been fiddling with the settings for
> ages, and it's still blotchey.
Your error_bound needs to be lowered to something like 0.1 probably; higher
values can look okay sometimes and go faster but blotches can appear
readily, I think. The recursion_limit 5 might might be overdoing it, not
sure. I tend to keep to less than that, but for speed. And I also usually
use a lesser nearest_count, but these things either depend on the conditions
of the scene or personal preference.
> Oh, and about that pretrace step... It seems to interact with the normal
> mosaic preview INI setting... Is this correct? Does changing the INI
> settings affect the final image, or only the preview process?
If I remember right, the mosaic preview is only a visual thing and the
pretrace does the refining or sampling control.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3dd56a62@news.povray.org>,
"Andrew Coppin" <orp### [at] btinternetcom> wrote:
> As I understand it, anything with a non-zero ambient setting will in effect
> "glow". My experiments seem to back this up.
This is true.
> Only trouble is, they don't glow very much; I'm having to use things
> like "ambient 10" to get enough light in my scene! (That's fine if
> the thing is white - doesn't work too well for pastel shades though.)
That isn't a problem. If you think about it, the intensity of light at
the surface of a fluorescent tube is much higher than the intensity of
light reflecting from a sheet of white paper. These very bright values
are necessary and correct.
I don't know what you mean about it not working for pastel shades,
though.
> Also, the images are very patchy. Now, when I just dunked an empty
> radiosity{} block into my scene, I wasn't suprised that the defaults didn't
> magically work perfectly. But... I've been fiddling with the settings for
> ages, and it's still blotchey. I've managed to improve it quite a bit, but
> it's still there. POV-Ray is obviously trying its hardest to smudge out
> those pixel artifacts from the pretrace step... but now quite suceeding.
Well, it sounds like you are trying to light the scene entirely with
radiosity. This is possible, but it usually takes a bit of experience to
get right...you are starting out with one of the more difficult ways to
use radiosity.
There was a craze of radiosity lit images on these newsgroups a while,
try searching the *.binaries.images group for radiosity images.
--
Christopher James Huff <cja### [at] earthlinknet>
http://home.earthlink.net/~cjameshuff/
POV-Ray TAG: chr### [at] tagpovrayorg
http://tag.povray.org/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Christopher James Huff wrote:
> I don't know what you mean about it not working for pastel shades,
> though.
He probably means the clipping of colors to <1,1,1>, so _any_ color with a
sufficient ambient value will look pure white. Kari's exposure simulation
patch solves this problem too quite nicely.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
If you don't need oddly-shaped area light sources, I really recommend you
using light_sources (point or area_lights, whichever is best in the scene).
They will help you making the scene render faster and look better. As a
bonus you also can get highlights on objects and even photons if you need
them.
--
#macro N(D)#if(D>99)cylinder{M()#local D=div(D,104);M().5,2pigment{rgb M()}}
N(D)#end#end#macro M()<mod(D,13)-6mod(div(D,13)8)-3,10>#end blob{
N(11117333955)N(4254934330)N(3900569407)N(7382340)N(3358)N(970)}// - Warp -
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Johannes Dahlstrom wrote:
> _any_ color
Well, any color apart from pitch black, that is.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Andrew Coppin wrote:
>
> I'm having to use things like "ambient 10" to get enough
> light in my scene!
Let's say you have a light source simulating a diffuse patch
like in cornell.pov. To make an equivalent ambient source use the
color of the light source multiplied by 2*pi as the object color
with ambient 1 (diffuse 0). Or separate the color and put the
multiplier in ambient term. For the cornell scene that would be 44
so high ambient values are normal.
For a simple test scene these settings will produce roughly equal
quality. Basically use higher count and nearest_count for ambient
sources. These particular settings may not work for your scene.
radiosity{
pretrace_start 0.04
pretrace_end 0.01
error_bound 0.4
#local RecursionLimit=2;
#if(AMBIENT_SOURCE)
count 1600
nearest_count 3
recursion_limit RecursionLimit+1
#else // for equivalent light_source
count 200
nearest_count 1
recursion_limit RecursionLimit
#end
}
_____________
Kari Kivisalo
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>> Any help would be appreciated. I can post an image if it helps - but
>> basically it's a large empty room with a smallish opening in the roof
that
>> lets in the light from a superambient object outside.
>> [...]
>
>This is never a good idea, you will need an extremely high count (probably
>larger than the maximum of 1600) to get smooth results in such a
>situation. A high quality area light instead of the ambient object is
>usually a better solution.
>
>Christoph
So... in short, you can't use radiosity. (Unless it's on a scene where it
won't make any real difference anyway.)
Actually, I like the idea of using an area light instead. (There is a large
white object directly under the skylight that was supposed to throw the
light around the room - but it doesn't anyway!) I had actually noticed that
specular highlight doesn't work with radiosity. (I would have expected them
to work really, but I guess I can see why they doesn't.)
Any yes - that bit about "pastel shapes" was the fact that, for example, an
orange light like rgb <1, 0.5, 0> will come out yellow with ambient 10. (And
yet - I presume! - the room will be lit up orange, not yellow. Haven't
actually tried it mind you ;-)
Oh well, thanx anyway people!
Andrew.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Andrew Coppin wrote:
>
> [...]
>
> So... in short, you can't use radiosity. (Unless it's on a scene where it
> won't make any real difference anyway.)
I don't understand. Why do you think radiosity makes no difference in a
scene with light sources?
The scene without light sources and only radiosity lighting is quite a
speciality. It works very well with hemispherical illumination or similar
situations (as you can find in the sample scenes and documentation) but
like i already explained small objects with high emission of light are
difficult. I have never seen a scene with this working well, neither with
POV-Ray nor with any other rendering package.
Christoph
--
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 15 Nov. 2002 _____./\/^>_*_<^\/\.______
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
