Well I finally gave in and added a light source to this image. I tried for
about a week to render this using high quality settings but everything I
tried made the artefacts worse. At an error_bound of 0.08 it starts to look
extremely mottled, though the artefacts are definitely smaller. I could
probably get rid of them completely by taking the error_bound below 0.01 and
if I didn't mind spending a month rendering the sucker. I think I might make
a poster out of this version though. The original was rendered at 1280x1024.

Post a reply to this message

Attachments:
Download 'trail.jpg' (88 KB)

Preview of image 'trail.jpg'

I liked the original - but agree the cylinder was crying out for
caustics. The only light source has to be from above so I don't think
your caustics will be brilliant. Overall, this one looks more like a
water-colour painting, or damp concrete walls. I still like it tho'.
Never did have the patience myself to play with rad!

Alf


"Thomas Lake" <tla### [at] homecom> wrote in message
news:3b5b4238@news.povray.org...
> Well I finally gave in and added a light source to this image. I
tried for

Post a reply to this message

I like the concept, but those radiosity artifacts look really terrible. If
it were me, I would introduce a surface normal on the walls of the
structure.. something subtle might get rid of the problem, yet still look
fairly smooth.

Andy


"Thomas Lake" <tla### [at] homecom> wrote in message
news:3b5b4238@news.povray.org...
> Well I finally gave in and added a light source to this image. I tried for
> about a week to render this using high quality settings but everything I
> tried made the artefacts worse. At an error_bound of 0.08 it starts to
look
> extremely mottled, though the artefacts are definitely smaller. I could
> probably get rid of them completely by taking the error_bound below 0.01
and
> if I didn't mind spending a month rendering the sucker. I think I might
make
> a poster out of this version though. The original was rendered at
1280x1024.
>
>
>

Post a reply to this message

> I like the concept,

Thanks

>but those radiosity artifacts look really terrible.

I definetly agree

>If it were me, I would introduce a surface normal on the walls of the
> structure.. something subtle might get rid of the problem, yet still look
> fairly smooth.

I could do that but it would really be dissapointing if I had to, a good
part of what makes the image nice are the smooth walls.
>
> Andy
>
>
> "Thomas Lake" <tla### [at] homecom> wrote in message
> news:3b5b4238@news.povray.org...
> > Well I finally gave in and added a light source to this image. I tried
for
> > about a week to render this using high quality settings but everything I
> > tried made the artefacts worse. At an error_bound of 0.08 it starts to
> look
> > extremely mottled, though the artefacts are definitely smaller. I could
> > probably get rid of them completely by taking the error_bound below 0.01
> and
> > if I didn't mind spending a month rendering the sucker. I think I might
> make
> > a poster out of this version though. The original was rendered at
> 1280x1024.
> >
> >
> >
>
>

Post a reply to this message

> I liked the original

As far as the rad artifats go I agree with you. But actually after adding
the light source the actual mood of this version is better than the last
IMO.

>- but agree the cylinder was crying out for
> caustics.

I used photons not Pov's regular cautics.

>The only light source has to be from above so I don't think
> your caustics will be brilliant. Overall, this one looks more like a
> water-colour painting, or damp concrete walls. I still like it tho'.
> Never did have the patience myself to play with rad!

Yeah it can really eat up a lot of your time! I was wondering if I am
experiencing the same slowdown that some of the others have commented on
with the combo of area light + ior + rad??

>
> Alf

Post a reply to this message

Thomas Lake wrote:

> I could do that but it would really be dissapointing if I had to, a good
> part of what makes the image nice are the smooth walls.

I think Alf is right though. I'm under the impression that it's quite impossible
to obtain both artefact-free walls and sharp shadows in the corners when using
the present implementation of Megapov's radiosity, at least for those who'd like
to see the finished render in their lifetime ;-).
However, when using irregular patterns on the walls (or even image maps), the
problem tends to disappear (read: to become less noticeable) and the results may
be fairly good.

G.

--

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

Post a reply to this message

"Thomas Lake" <tla### [at] homecom> wrote in message
news:3b5bb552$1@news.povray.org...
> > Never did have the patience myself to play with rad!
>
> Yeah it can really eat up a lot of your time! I was wondering if I am
> experiencing the same slowdown that some of the others have commented on
> with the combo of area light + ior + rad??

Heck, just point light source + ior + radiosity in a scene of a screened
room I'm doing right now is increasing render-time much more than I would
have expected.  I have caustics 1 used here, not photons (yet).
Actually, I had other objects (semitransparent people for scale) using ior
but what caused the slowdown is a several roof panels (corrugated
fiberglass) with a 2-layer texture.  I think it's more related to AA of
those in my case.

Bob H.

Post a reply to this message

> I think Alf is right though. I'm under the impression that it's quite
impossible
> to obtain both artefact-free walls and sharp shadows in the corners when
using
> the present implementation of Megapov's radiosity, at least for those
who'd like
> to see the finished render in their lifetime ;-).
> However, when using irregular patterns on the walls (or even image maps),
the
> problem tends to disappear (read: to become less noticeable) and the
results may
> be fairly good.

Really? Hmm I'll have to give it a try then, thanks Gilles.

Post a reply to this message