Help, gurus!

I have several indoor scenes where the primary source of illumination is
sunlight coming through one (or several windows).  After playing with
gigantic white spheres outside my building, parallel lights coming through
the windows, and every radiosity setting combination I can think of, I
cannot get the room to look real.  Some typical results:

- If I crank my "sunlight" (a parallel light) up to rgb<100,100,100> I get
very blotchy results, though the room is lit reasonably well
- If I crank my "ambient white sphere" (a gigantic sphere outside my
building) to have ambient 100, I get strange artifacts
- If I leave lighting values down where they are reasonable (2 being a
maximal value) my room is too dark, the light doesn't bounce in the room at
all
- If I turn recursion up to 20 and increase my count to 500, my results get
a bit better... but it is REALLY slow and still very dark
- Using a light INSIDE my room gives me great results!  It is just the
sunlight that I am having trouble with.

Some notes:

- My materials all have a diffuse value from 0.9 to 1.0 - so they should be
scattering the light around the room nicely
- I tried playing with "Brightness" above values of 1.0 which light up my
room, but make my contrast levels look strange

... my goal here is to have a generic "sunlight" method that I can use with
several indoor scenes, which does not require an hour for each frame.  Is
this a feasible task?  Please tell me that I don't need to add area lights
to each window.  :)

Any suggestions greatly appreciated, sample scenes, anything.

Cheers,
JP

Post a reply to this message

Hi John,

I haven't tried myself, but JRG seems to know the trick and he just
explained it in p.b.i. in the tread "ACE".. Here's a direct link:

http://news.povray.org/povray.binaries.images/25857/?mtop=183588&moff=10


Regards,
Hugo

Post a reply to this message

John Pallett wrote:
>
> my goal here is to have a generic "sunlight" method that I can use with
> several indoor scenes

Here is one which uses realistic energies for sun and sky.
The net diffuse value for concrete is realistic 0.4.
http://luxlab.com/tmp/balance.jpg

Same method applied to a model where the materials of the
outside structures were at some random default values resulting
from multiple conversions. The inside materials were ok (net diffuse
0.25 for sandstone) so it's mostly realistic.

http://luxlab.com/tmp/church2.jpg


> which does not require an hour for each frame

Very high quality radiosity was used in the first scene because
measurements from the color probes needed to be accurate. Lower
quality settings give faster but less accurate results. It's still
going to be slow and difficult to use. If interested mail me.


_____________
Kari Kivisalo

Post a reply to this message

It's no doubt something that can't be described easily, it kind of sounds
like you expect radiosity to be a light source.

When you say "too dark" that means you aren't seeing a brightness you
expect, something which others might not expect to see in a sunlit room with
no other light sources except for the Sun through a window. If the room has
dark texturing it will of course still be darkened even though sunlight
shines in, and if it is a bright white interior it could almost glow in
places where the sunlight shines. However, it's a visual thing. The eyes
adjust to it in various ways. Try the following scene and see what you
think; you'll probably need to show a rendering of what you don't want and
describe how it should be altered, or make changes to it to show how it
ought to look.
You mentioned diffuse values up to 1 but not over one. Perfectly legal to
use diffuse 1.1 and beyond. I'm thinking that the problem there is how the
radiosity is going to blend everything together. It usually shows most at
very short distances and tapers off abruptly. I'm not sure of how to go
about counteracting that effect and instead get a uniform, farther reaching
effect.

#version 3.5;

global_settings {
        radiosity {
                brightness 0.9
                recursion_limit 3
                gray_threshold 0.3
                media on
        }
}

// ----------------------------------------

camera {
  location  <-1, 2, -6>
  right     x*image_width/image_height
  look_at   <0,0,0>
}

// sky
sphere {
        0,1
  pigment {
    gradient y
    color_map {
      [0 rgb <0.8,0.9,1.0>]
      [0.7 rgb <0.3,0.5,0.9>]
      [1 rgb <0.2,0.4,0.7>]
    }
  }
  finish {ambient 0.9 diffuse 0.1}
  scale 9999
 hollow
}

// sunlight
light_source {
        0,<1.5,1.3,1.1>
        looks_like {
                sphere {0,99 pigment {rgb 1} finish {ambient 1 diffuse 1}}
                }
        parallel
        point_at z
 translate -999*z rotate <30,45,0>
}

// ground
plane {
  y, -0.01
  pigment { color rgb <0.3,0.7,0.4> }
}

// ----------------------------------------

// visibility test objects
#local W=5;
#while (W>-5)
sphere {
  0,0.5
  translate W*z
        pigment {checker color rgb 0 color rgb 1 scale 0.25}
        finish {ambient 0.1 diffuse 0.6} // defaults
}
#local W=W-1;
#end

// room with 4 windows
difference {
box {
        <-3,0,-9>,<3,3,3>
}
box {
        <-3,0.1,-9>/1.1,<3,3,3>/1.1
}
box {
        <-3.1,1,-2>,<3.1,2,-1>
}
box {
        <-3.1,1,1>,<3.1,2,2>
}

        pigment {rgb 0.9}
        finish {ambient 0.25 diffuse 0.75 specular 0.1 roughness 0.1}
 hollow
}

// air in room
box {
        <-3,0,-9>,<3,3,3>
        pigment {rgbf 1}
        interior {
                media {
                        emission 0.6
                        absorption 0.9
                        scattering {
                                5,3
                                eccentricity 0.3 // use with type 5
                                extinction 0
                        }
                        density {
                                rgb 0.1
                        }
                        samples 30,60
                }
        }
 hollow
}

Post a reply to this message

John Pallett <joh### [at] chiefarchitectcom> wrote:
> - If I crank my "sunlight" (a parallel light)

  Btw, sunlight is not parallel.

-- 
#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

In article <3d417105@news.povray.org>, Warp <war### [at] tagpovrayorg> 
wrote:

> John Pallett <joh### [at] chiefarchitectcom> wrote:
> > - If I crank my "sunlight" (a parallel light)
> 
>   Btw, sunlight is not parallel.

It's close enough for most purposes...

-- 
Christopher James Huff <chr### [at] maccom>
POV-Ray TAG e-mail: chr### [at] tagpovrayorg
TAG web site: http://tag.povray.org/

Post a reply to this message

Try setting assumed_gamma to somthing very low, like 1 or 0.5. This will 
brighten everything up, hopfully simulating the natural adjustment our 
eyes make in low-light situations.

John Pallett wrote:

> Help, gurus!
> 
> I have several indoor scenes where the primary source of illumination is
> sunlight coming through one (or several windows).  After playing with
> gigantic white spheres outside my building, parallel lights coming through
> the windows, and every radiosity setting combination I can think of, I
> cannot get the room to look real.  Some typical results:
> 
> - If I crank my "sunlight" (a parallel light) up to rgb<100,100,100> I get
> very blotchy results, though the room is lit reasonably well
> - If I crank my "ambient white sphere" (a gigantic sphere outside my
> building) to have ambient 100, I get strange artifacts
> - If I leave lighting values down where they are reasonable (2 being a
> maximal value) my room is too dark, the light doesn't bounce in the room at
> all
> - If I turn recursion up to 20 and increase my count to 500, my results get
> a bit better... but it is REALLY slow and still very dark
> - Using a light INSIDE my room gives me great results!  It is just the
> sunlight that I am having trouble with.
> 
> Some notes:
> 
> - My materials all have a diffuse value from 0.9 to 1.0 - so they should be
> scattering the light around the room nicely
> - I tried playing with "Brightness" above values of 1.0 which light up my
> room, but make my contrast levels look strange
> 
> ... my goal here is to have a generic "sunlight" method that I can use with
> several indoor scenes, which does not require an hour for each frame.  Is
> this a feasible task?  Please tell me that I don't need to add area lights
> to each window.  :)
> 
> Any suggestions greatly appreciated, sample scenes, anything.
> 
> Cheers,
> JP
> 
> 
> 


-- 
Samuel Benge

sbe### [at] caltelcom

Post a reply to this message

In article <3D4### [at] caltelcom>,
 Samuel Benge <sbe### [at] caltelcom> wrote:

> Try setting assumed_gamma to somthing very low, like 1 or 0.5. This will 
> brighten everything up, hopfully simulating the natural adjustment our 
> eyes make in low-light situations.

Very low? 1 is the recommended value..."For new scenes, you should use 
an assumed gamma value of 1.0 as this models how light appears in the 
real world more realistically."

-- 
Christopher James Huff <chr### [at] maccom>
POV-Ray TAG e-mail: chr### [at] tagpovrayorg
TAG web site: http://tag.povray.org/

Post a reply to this message

> Any suggestions greatly appreciated, sample scenes, anything.

Just an idea: You ought to need a realistic sky for a realistic scene. At
least, the sun shouldn't be a white sphere (or spherical pattern) in a black
sky, but in a blue-to-white sky. This will lighten up your room a bit...

cukk

Post a reply to this message

Hi guys -

Some more notes -

I was surrounding my entire scene with a gigantic white sphere, not creating
a 'spherical sun'.  The problem, as it turns out, was that the light wasn't
getting scattered around the room enough for it to properly simulate
sunlight coming in a window (which scatters all over the room).  So the
scene consisted of:

- A huge white sphere, ambient 0.5
- A parallel white light (sunlight) (parallel lighting for the sun is good
enough for our purposes)
- A room with a window, all materials set to (ambient 0, diffuse 0.9) so
that they aren't emitting light into the scene
- My camera inside the room

Thanks to everyone for suggestions thus far.  The missing ingredient seems
to have been media, which can be used to scatter light.  What I am doing now
is using media inside my rooms to scatter the sunlight as it comes in the
window and as it bounces around the room.  Very subtle scattering gives me
great results.

I'll post images to PBI once I get something put together that demonstrates
the problem clearly.

Can anyone suggest another way to solve the problem of having the light
coming in the window scatter around the room appropriately?  Note that my
camera points TOWARDS the window, so making the window's wall invisible is
not an option.

JP

Post a reply to this message