POV-Ray : Newsgroups : povray.newusers : Getting good glows over high-intensity areas Server Time
29 Jul 2024 20:25:42 EDT (-0400)
  Getting good glows over high-intensity areas (Message 11 to 16 of 16)  
<<< Previous 10 Messages Goto Initial 10 Messages
From: Bob Hughes
Subject: Re: area lights
Date: 20 Jul 2005 12:07:41
Message: <42de76cd$1@news.povray.org>
"ZainAnak" <nomail@nomail> wrote in message 
news:web.42de5e3fa93ad1b23b6b98ba0@news.povray.org...
> Bob, thanks for all the help!

Always welcome to any I have.

> sort of area light setup?  All I really want, after all, is the sun lines
> to be a *little* blurrier than normal- I don't want everything bright in
> the scene to glare out, like your sample code would do.  Here is my 
> current
> light:

 light_source {
  <-5000, 4000, 15000>
  color rgb <0.8, 0.4, 0.2>*4.1
  area_light 150*x, 150*y, // dimensions, or physical size
  4, 4 // number of sources for each axis (above). integers only
  adaptive 1 // sampling rays (more is slower, but better. integers only)
  jitter // smooth shadows, without this you would get banding
  circular // make it round not square
  orient // turn toward scene objects (works in conjunction with circular)
 }

Ha ha ha! Well, you're right. Why not keep it simple? Heck, I'm just glad 
you got me going on Samuel's light-bleed idea, otherwise I might have never 
had such a thing here.

When you're dealing with area_light the important things are size of the 
array and smoothness of the shadows. I've added in what might be okay with 
you're scene but you'll definitely need to adjust the x and y dimensions to 
either narrow or widen the resulting shadows.
The number of sources (4 X 4) ultimately determines quality, as well as 
speed of the render, so you might also need to increase that even if it 
slows the rendering down. But if the shadow edges are only needed to be very 
thin you could possibly lower those two numbers (equally and integer only) 
to a 3 by 3 array. All depends on what you see in the render as you make 
adjustments.

Oh yeah, and area_light does not affect surfaces directly, such as specular 
highlights and the diffuse finishes. It's only for shadows.

You would be wise to read up on this in the scene help (docs), section 
3.4.7.5  Area Lights, so I'll leave it at that. Unless you have more to ask, 
please go ahead and do so.  : )

Bob


Post a reply to this message

From: ZainAnak
Subject: Re: area lights
Date: 20 Jul 2005 14:25:00
Message: <web.42de965ef335b5b53b6b98ba0@news.povray.org>
Bob, this is looking GREAT.  I'm futzing around with the settings, and it's
giving me some nice color ramps where the sun falls.  And I don't even have
to render twice... woohoo!

Thanks!
-ZA


Post a reply to this message

From: Slime
Subject: Re: area lights
Date: 20 Jul 2005 17:16:14
Message: <42debf1e$1@news.povray.org>
>   4, 4 // number of sources for each axis (above). integers only

So low! I generally use 17, 17 myself. Maybe I could go lower if I used
jitter more often, but I still recommend using higher numbers than 4; the
jittering with such a low number probably forces otherwise unnecessary
anti-aliasing.

Of course, with such high numbers it's very important to use adaptive (so
that you don't actually shoot a full 17x17=289 shadow rays); adaptive 1 has
always worked best for me.

Finally, I also recommend using a number which is one greater than a power
of 2 (3, 5, 9, 17, 33) because these make the most sense with the adaptive
process (See the second picture at
http://www.povray.org/documentation/view/3.6.1/313/ , which demonstrates why
this makes sense).

 - Slime
 [ http://www.slimeland.com/ ]


Post a reply to this message

From: Bob Hughes
Subject: Re: area lights
Date: 20 Jul 2005 18:29:59
Message: <42ded067$1@news.povray.org>
"Slime" <fak### [at] emailaddress> wrote in message 
news:42debf1e$1@news.povray.org...
>>   4, 4 // number of sources for each axis (above). integers only
>
> So low! I generally use 17, 17 myself. Maybe I could go lower if I used
> jitter more often, but I still recommend using higher numbers than 4; the
> jittering with such a low number probably forces otherwise unnecessary
> anti-aliasing.

To be honest, I actually wrote 5,5 at first and replaced it with 4,4 
thinking it might seem too slow. I subconsciously consider shorter render 
times to be better than quality of the output. A habit formed during the 
days of one-line-at-a-time renderings and hardly ever doing "final" images 
left running for hours or days.

> Of course, with such high numbers it's very important to use adaptive (so
> that you don't actually shoot a full 17x17=289 shadow rays); adaptive 1 
> has
> always worked best for me.

Makes sense. Proves you *can* teach an old dog new tricks. I've always 
stayed away from the larger arrays just because I worried about longer 
render times.

> Finally, I also recommend using a number which is one greater than a power
> of 2 (3, 5, 9, 17, 33) because these make the most sense with the adaptive
> process (See the second picture at
> http://www.povray.org/documentation/view/3.6.1/313/ , which demonstrates 
> why
> this makes sense).

I went looking for the example files and read each lights\arealight*.pov, as 
well as rendering them in a few different ways to check against my usual way 
and what you said. Firstly, I hadn't realized those had 17 by 17 arrays. 
Secondly, I'm amazed to find out that circular.pov has only 4 by 4 arrays 
and the shadows are awful. But even more amazing is that after changing them 
to 9 by 9 the render time only increased from 11.56 to 11.64 seconds! The 
shadows look great, too; although, any higher than 5 by 5 array and the idea 
behind the example of a circular arealight compared to non-circular gets 
lost. ;-)

Apparently I'm going to need to change my mind about arealights and start 
giving better advice about them in the future.

Thankyou for jostling my old POVing into new territory! If I can just train 
myself not to drop back to 4X4's again I should be okay.

Bob


Post a reply to this message

From: Artur Rataj
Subject: Re: more to be said
Date: 24 Jul 2005 10:25:01
Message: <web.42e3a3d9a93ad1b2e9d8b2350@news.povray.org>
There might be one more effect in the picture, but the ones you discussed in
this thread, that would cause the line to be blurred: if the wall is
bumped, the bumps in the wall may reflect light, causing the neighboring
bumps to be lit, even that they are not directly lit by the sun. Radiosity
or photon maps might be used for this.

"ZainAnak" <nomail@nomail> wrote:

> I had an idea, though.  If you look at the original picture I posted as a
> reference, you'll notice that it's a pretty small, well-defined bleed that
> occurs at the sun line.  I was wondering- couldn't I achieve this with some
> sort of area light setup?  All I really want, after all, is the sun lines
> to be a *little* blurrier than normal- I don't want everything bright in
> the scene to glare out, like your sample code would do.  Here is my current
> light:
>
> light_source {
>   <-5000, 4000, 15000>
>   color rgb <0.8, 0.4, 0.2>*4.1
> }
>
> how would you turn this into a light with slightly fuzzier lines?
>
> Thanks!
> -ZA


Post a reply to this message

From: Artur Rataj
Subject: Re: more to be said
Date: 24 Jul 2005 10:45:00
Message: <web.42e3a829a93ad1b2e9d8b2350@news.povray.org>
And there might be yet one more effect: the bumps might cause the shadow
line to `wiggle', what may look like a kind of softening from some
distance.


Post a reply to this message

<<< Previous 10 Messages Goto Initial 10 Messages

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.