The docs say @ 10.2.3. "distance_maximum is automatically computed"

I have an outdoor nighttime scene with lots of lit buildings. They are
far apart.  I want the lighting interior to the buildings to light up
everything, including outdoors and outsides of next buildings.  I think
radiosity will be a neat trick, avoiding having 80 light sources (my PC
only has 64MB RAM!).

My first tweaks at coding suggest that there's not much "interior light"
making it to the next building.    Any ways around this? Is it possible
that the change above 10.2.3 made some rooms better but limited
applicability of radiosity to certain scenarios? Sorry if that's an
ornery question......


SECOND QUESTION:  is it true that point light sources will not interact
with media:  only spotlights will do so?  The scenario here is "point
light source inside a wiffle ball."  If I set up some media, will I see
streaks of light pouring out the wiffle ball holes?

THANKS!

Post a reply to this message

"Greg M. Johnson" wrote:

> My first tweaks at coding suggest that there's not much "interior light"
> making it to the next building.    Any ways around this? Is it possible
> that the change above 10.2.3 made some rooms better but limited
> applicability of radiosity to certain scenarios?

This is not so strange. For radiosity to work well, it's better to have
large areas of bright surfaces. In night scenes, where the amount of
available light is limited, radiosity effects may not be very noticeable.
I've tried several dark scenes where radiosity didn't make any difference
and where secondary or fill lights gave much prettier results. Of course it
can make a big difference if the scene shows what's inside the building, or
if you have a colored sky_sphere, or if you have direct outdoor lights, or
if you have a very bright building etc. You can try increasing the radiosity
brightness too.

About your 80 lights, just turn off the light buffers with -UL on the
command line. The rendering time will be slightly higher, but you won't run
into memory problems.

> SECOND QUESTION:  is it true that point light sources will not interact
> with media:  only spotlights will do so?

No. It's just that spotlights render faster because (if I have understood
well) the calculations are limited to the beam. Your wiffle ball should work
OK with a pointlight. Using several spotlights could perhaps make it faster,
but that remains to be tested.

G.

--

**********************
http://www.oyonale.com
**********************
Graphic experiments
Pov-ray gallery

Post a reply to this message

In article <3A6831F8.99A66548@my-dejanews.com>, 
gre### [at] my-dejanewscom wrote:

> My first tweaks at coding suggest that there's not much "interior light"
> making it to the next building.    Any ways around this? Is it possible
> that the change above 10.2.3 made some rooms better but limited
> applicability of radiosity to certain scenarios? Sorry if that's an
> ornery question......

Try raising the recursion level...this will increase the illumination in 
areas that are illuminated by light that has gone through a couple 
bounces. Also make sure enough samples are being used, and maybe raise 
the "brightness" value. Maybe you should add a full moon. (Don't rely on 
radiosity from it though, it's too small...use a point light.)


> SECOND QUESTION:  is it true that point light sources will not interact
> with media:  only spotlights will do so? 

That's not true at all...where did you hear that? All lights can 
interact with media.


> The scenario here is "point light source inside a wiffle ball."  If I 
> set up some media, will I see streaks of light pouring out the wiffle 
> ball holes?

If you use scattering media and if media_interaction for the 
light_source isn't turned off, then yes. (Assuming media density and 
light brightness are set correctly, of course)

-- 
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

"Greg M. Johnson" <gre### [at] my-dejanewscom> wrote in message
news:3A6831F8.99A66548@my-dejanews.com...
>
> My first tweaks at coding suggest that there's not much "interior light"
> making it to the next building.    Any ways around this? Is it possible
> that the change above 10.2.3 made some rooms better but limited
> applicability of radiosity to certain scenarios? Sorry if that's an
> ornery question......
>

Funnily enough, I've been playing with this problem and have a solution that
seems to work. It's a kludge and will no doubt offend any radiosity purists,
but what the hell.

Add a low "assumed_gamma" setting to your global block (1/2 seems pretty
good). For an example, try rendering the following code:

#version unofficial MegaPov 0.6;

#include "colors.inc"
global_settings{
  assumed_gamma 1/2
  ambient_light 0
  ini_option "+QR"
  radiosity{
    pretrace_start 0.08
    pretrace_end   0.02
    count 150 //300
    nearest_count 10
    error_bound 1
    recursion_limit 2
    low_error_factor 1/2
    gray_threshold 0.5
    minimum_reuse 0.015
    brightness 1
    adc_bailout 0.01/1
  }
}

camera{location  <35.0, 60.0, -100.0> look_at   <0.0, 10.0,  0.0>}
light_source{ <0,0,0> color rgb<1,1,1/2>*0.75  translate y*35}
difference{
  cylinder{0, y*50, 10}
  cylinder{0, y*50, 10 scale 0.9}
  cylinder{y*25, <12,25,0>, 2}
  pigment{White}
}
cylinder{0, y*50, 10 translate x*40 pigment{White}}
cylinder{0, y*50, 10 translate z*50 pigment{White}}
sphere{0,10 translate<-20,10,-30> pigment{White}}
plane{y, 0 pigment{White}}

Post a reply to this message