Hi Folks,

This is a simple test scene to see how well PoV handles radiosity 
lighting by long thin ambient sources - effectively fluorescent strip 
lighting.

The two images are

room_orig.jpg - uses standard WinPoV 3.5
room_mod.jpg  - uses a radiosity modified personal compile of 3.5

Both use scene posted in p.b.s-f.

See the thread in p.a-u as well.

Comments please?

Bye for now,
     Mike Andrews.

Post a reply to this message

Attachments:
Download 'room_orig.jpg' (24 KB) Download 'room_mod.jpg' (22 KB)

Preview of image 'room_orig.jpg'

Preview of image 'room_mod.jpg'

"Michael Andrews" <m.c### [at] readingacuk> wrote in message
news:3de615f6@news.povray.org...
>
> This is a simple test scene to see how well PoV handles radiosity
> lighting by long thin ambient sources - effectively fluorescent strip
> lighting.

Hmmm. Those are very different alright. However... I just tried a test run
of your scene script and using a count of 400 (much faster than at 1600) and
error_bound 1.0 it looks okay too. So, guess what I'm wondering is, how is
your patched POV better than using other values in the official POV? I
realize of course that the settings I used might not work out for other
renders, plus I only checked with a tiny resolution to see it quicker.

Post a reply to this message

What were the running times on each? (the second one does look much better)

Pentahall (formerly quadhall, formerly treshall)


"Michael Andrews" <m.c### [at] readingacuk> wrote in message
news:3de615f6@news.povray.org...
> Hi Folks,
>
> This is a simple test scene to see how well PoV handles radiosity
> lighting by long thin ambient sources - effectively fluorescent strip
> lighting.
>
> The two images are
>
> room_orig.jpg - uses standard WinPoV 3.5
> room_mod.jpg  - uses a radiosity modified personal compile of 3.5
>
> Both use scene posted in p.b.s-f.
>
> See the thread in p.a-u as well.
>
> Comments please?
>
> Bye for now,
>      Mike Andrews.
>
>
>


----------------------------------------------------------------------------
----






----------------------------------------------------------------------------
----

Post a reply to this message

This looks promising! I read your post in advanced-users, but I'm not sure I
understand the sideeffect of your patch. Could you post an image where it
goes wrong (I know.. strange request) and more images to demonstrate how
well it behaves. What about spherical light sources? ... I won't set my
hopes too high, but on the other hand... :o)

Best regards,
Hugo

Post a reply to this message

hughes, b. wrote:
> Hmmm. Those are very different alright. However... I just tried a test run
> of your scene script and using a count of 400 (much faster than at 1600) and
> error_bound 1.0 it looks okay too. So, guess what I'm wondering is, how is
> your patched POV better than using other values in the official POV? I
> realize of course that the settings I used might not work out for other
> renders, plus I only checked with a tiny resolution to see it quicker.

Using a high error_bound will smooth out a lot of the noise, but also 
smears the shadows.

I've attached a cropped piece of the same scene run with

   radiosity {
     pretrace_start 16/image_width
     pretrace_end   2/image_width
     count 400
     nearest_count 15
     error_bound 1
     recursion_limit 2
     low_error_factor .2
     gray_threshold 0.0
     minimum_reuse 0.001
     brightness 1
     adc_bailout 0.01/30
     max_sample 30
     normal on
   }

The flat surfaces look nice and smooth, but compare the shadows around 
the chairs to those in the previous images I posted - you'll see that 
the lower error_bound gives a sharp shaddow edge which you would expect 
to get from a light like this.

Mike Andrews.

Post a reply to this message

Errrm ... and now the image ... 8-/

Mike.

Post a reply to this message

Attachments:
Download 'room_high_ef.jpg' (4 KB)

Preview of image 'room_high_ef.jpg'

Okay, thanks. Yeah, that shows an improvement as well. As did the first
message posting did already, but I had to try and be a skeptic. Sure hope
your fix proves good all around. And, if so, that it can be incorporated
into POV-Ray.

Post a reply to this message

hughes, b. wrote:
> Okay, thanks. Yeah, that shows an improvement as well. As did the first
> message posting did already, but I had to try and be a skeptic. Sure hope
> your fix proves good all around. And, if so, that it can be incorporated
> into POV-Ray.

Thanks. It's not really a 'fix' though because the radiosity code isn't 
really 'broken'. It's just that under some specific conditions I've 
tried (ie small radiosity light sources) it can give an improvement.

I'll try to get a proper patch together at the weekend.

Mike Andrews.

Post a reply to this message