I put together renders of that sample chess scene file that I has been around
since at least POV-Ray 2.2, compared with the chess2.pov of version 3.1g and
later.
Realized I must have added the missing(?) 2nd ear on the knight/horse before the
release of version 3.5. Never can remember if the single ear was intentional by
chess.pov scene creator Ville Saari.

Been slowly making changes to that scene file the past several months, turning
the chess set into an include file trimmed down for inclusion with any other
scene. Finally have both the inc and pov files able to render both ways, the
original and modified.

Made changes and additions to the chess pieces so that they more closely match
sizes with a contemporary Staunton set. Most changes done to the knight/horse to
add a mane, mouth and eyes; along with tweaking the overall shape.

Main thing besides was to get rid of the quadric-based shapes from the now-named
shapes_old.inc file so that it could be bypassed entirely. I replaced them with
isosurfaces (have I mentioned how bad I am with those?) and conventional
primitives instead.

It's not mathematical perfection, or clean considering the rescaling and
translate changes I had to do, but it seems to have worked okay.

I will post the files to p.b.s-f. for your review. Please note that I was trying
features from the latest releases of 3.7.1.1-alpha and UberPOV but it should be
suitable for rendering with official 3.7. They are commented out.

The needed chess-set_mod.inc to follow this message.

Bob

Post a reply to this message

Attachments:
Download 'chess2-37.pov.txt' (17 KB)

"omniverse" <omn### [at] charternet> wrote:
> Please note that I was trying
> features from the latest releases of 3.7.1.1-alpha and UberPOV but it should be
> suitable for rendering with official 3.7. They are commented out.

Specular highlights used for Ivory_M and Ebony_M material/texture/finish are
probably too bright unless used in conjunction with fresnel, a feature of
current unofficial 3.7.1.1-alpha versions.

Bob

Post a reply to this message

Attachments:
Download 'chess-set_mod.inc.txt' (19 KB)

> "omniverse" <omn### [at] charternet> wrote:
>> Please note that I was trying
>> features from the latest releases of 3.7.1.1-alpha and UberPOV but it should be
>> suitable for rendering with official 3.7. They are commented out.
>
> Specular highlights used for Ivory_M and Ebony_M material/texture/finish are
> probably too bright unless used in conjunction with fresnel, a feature of
> current unofficial 3.7.1.1-alpha versions.
>
> Bob
>

Why does those finishes have any emission at all? Emission is only if 
you want a texture to actually glow, even if it's a very dim glow.
Beter remove all emission everywhere, or set it at zero everywhere.

Post a reply to this message

Alain <kua### [at] videotronca> wrote:

> > "omniverse" <omn### [at] charternet> wrote:
> >
> > Specular highlights used for Ivory_M and Ebony_M material/texture/finish are
> > probably too bright unless used in conjunction with fresnel, a feature of
> > current unofficial 3.7.1.1-alpha versions.
> >
>
> Why does those finishes have any emission at all? Emission is only if
> you want a texture to actually glow, even if it's a very dim glow.
> Beter remove all emission everywhere, or set it at zero everywhere.

Well, considering radiosity eliminates ambient lighting I just use a tiny bit of
emission to try and avoid any totally black areas. And also for a while now I've
been keeping ambient to much lower levels than the default 0.1 for non-radiosity
scenes, sometimes ambient 0 in lieu of a very low emission for the kind of
ambient lighting.

I'm used to the idea nothing is completely black unless in space or a cave, or
deep ocean. All of which could still have a very small amount of illumination
depending on the actual environs (glowing fish, phosphorescence) or perceived
unreal light at the eye. In other words I don't believe in total blackness.  :)

I don't expect anyone to use the exact scene as-is without making their own
changes too. I found the old scene warns about 'unusually high ambient' because
the original chess board uses ambient 0.35, and not only that they were using
assumed_gamma 2.2 so POV-Ray 3.7 really made them dark. Changing to
assumed_gamma 1.0 washes out the board.

Your question is very relevant, especially if considering this chess scene to
replace the current file in a new release of POV-Ray. And I did think that might
be a worthwhile thing to do later, but I wasn't going to ask yet. So for anyone
interested be my guest to discuss this further.

Bob

Post a reply to this message

Am 05.12.2016 um 06:42 schrieb omniverse:
> Alain <kua### [at] videotronca> wrote:

>>> "omniverse" <omn### [at] charternet> wrote:
>>>
>>> Specular highlights used for Ivory_M and Ebony_M material/texture/finish are
>>> probably too bright unless used in conjunction with fresnel, a feature of
>>> current unofficial 3.7.1.1-alpha versions.
>>>
>>
>> Why does those finishes have any emission at all? Emission is only if
>> you want a texture to actually glow, even if it's a very dim glow.
>> Beter remove all emission everywhere, or set it at zero everywhere.
> 
> Well, considering radiosity eliminates ambient lighting I just use a tiny bit of
> emission to try and avoid any totally black areas. And also for a while now I've
> been keeping ambient to much lower levels than the default 0.1 for non-radiosity
> scenes, sometimes ambient 0 in lieu of a very low emission for the kind of
> ambient lighting.

But avoiding completely black areas _is_ radiosity's _job_ (when it is
enabled).

In this manner it differs fundamentally from ambient occlusion, which
reduces the brightness of an image, especially where surfaces are close
together. Radiosity, in contrast, brightens up things, especially where
surfaces are far apart -- but it also does add some brightness to nooks
and crannies.

If you still see completely black areas, this can mean one of two (well,
maybe three) things:

(A) You have set `recursion_limit` too low. I'd recommend 1 or 2 for
wide open outdoor scenes, but 2 or 3 for indoor scenes. Add another
level if your scene has no conventional light sources and is illuminated
exclusively from sky spheres, emissive materials or emissive media.

(B) Your light sources are too bright and your materials too dark.
Without radiosity this cancels out and you'll never be able to notice,
but when radiosity enters the picture this makes nooks and crannies darker.

(C Using a high assumed_gamma will also cause nooks and crannies to look
darker, because the gamma's overall darkening effect is most pronounced
for very dark tones.)


An additional note to (A): When reaching the recursion limit, the
radiosity algorithm currently presumes everything to be pitch black;
i.e. if `recursion_limit 2` is set, the algorithm will pretend that
light rays can't bounce around between diffuse surfaces more than twice.
This has the effect that using a low recustion limit makes the entire
image a tad darker than would be realistic (the most realistic setting
would be infinity, but that's not realistically computable... pun intended).

I have plans in my drawer to extend the radiosity algorithm, and make it
respect "ambient" again -- but only once it reaches the recursion limit.
This should have the effect of brightening up the scene just enough to
approximately compensate for the lack of infinite recursion (provided
"ambient" is set to a reasonable value).

I might also add an ambient occlusion mode to the mix: If enabled, an
incomplete recursion step would be made, querying only distance
information, and using that to simply darken the effective "ambient"
component.

But that's future plans, so nobody hold their breath please...

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Am 05.12.2016 um 06:42 schrieb omniverse:
> > Alain <kua### [at] videotronca> wrote:

> >>> "omniverse" <omn### [at] charternet> wrote:
> >>>
> >>> Specular highlights used for Ivory_M and Ebony_M material/texture/finish are
> >>> probably too bright unless used in conjunction with fresnel, a feature of
> >>> current unofficial 3.7.1.1-alpha versions.
> >>>
> >>
> >> Why does those finishes have any emission at all? Emission is only if
> >> you want a texture to actually glow, even if it's a very dim glow.
> >> Beter remove all emission everywhere, or set it at zero everywhere.
> >
> > Well, considering radiosity eliminates ambient lighting I just use a tiny bit of
> > emission to try and avoid any totally black areas. And also for a while now I've
> > been keeping ambient to much lower levels than the default 0.1 for non-radiosity
> > scenes, sometimes ambient 0 in lieu of a very low emission for the kind of
> > ambient lighting.
>
> But avoiding completely black areas _is_ radiosity's _job_ (when it is
> enabled).
>
> In this manner it differs fundamentally from ambient occlusion, which
> reduces the brightness of an image, especially where surfaces are close
> together. Radiosity, in contrast, brightens up things, especially where
> surfaces are far apart -- but it also does add some brightness to nooks
> and crannies.
>
> If you still see completely black areas, this can mean one of two (well,
> maybe three) things:
>
> (A) You have set `recursion_limit` too low. I'd recommend 1 or 2 for
> wide open outdoor scenes, but 2 or 3 for indoor scenes. Add another
> level if your scene has no conventional light sources and is illuminated
> exclusively from sky spheres, emissive materials or emissive media.
>
> (B) Your light sources are too bright and your materials too dark.
> Without radiosity this cancels out and you'll never be able to notice,
> but when radiosity enters the picture this makes nooks and crannies darker.
>
> (C Using a high assumed_gamma will also cause nooks and crannies to look
> darker, because the gamma's overall darkening effect is most pronounced
> for very dark tones.)
>
>
> An additional note to (A): When reaching the recursion limit, the
> radiosity algorithm currently presumes everything to be pitch black;
> i.e. if `recursion_limit 2` is set, the algorithm will pretend that
> light rays can't bounce around between diffuse surfaces more than twice.
> This has the effect that using a low recustion limit makes the entire
> image a tad darker than would be realistic (the most realistic setting
> would be infinity, but that's not realistically computable... pun intended).
>
> I have plans in my drawer to extend the radiosity algorithm, and make it
> respect "ambient" again -- but only once it reaches the recursion limit.
> This should have the effect of brightening up the scene just enough to
> approximately compensate for the lack of infinite recursion (provided
> "ambient" is set to a reasonable value).
>
> I might also add an ambient occlusion mode to the mix: If enabled, an
> incomplete recursion step would be made, querying only distance
> information, and using that to simply darken the effective "ambient"
> component.
>
> But that's future plans, so nobody hold their breath please...

I can hold my breath forever... as long as no one is looking.  ;)

I never use more than a recursion of 2 (of course I would be lying there),
mostly 1 or not set. So I still shy away from 3, from past experiences with
render time shock. Especially when impatiently experimenting on a scene.

Sounds like I was under the wrong impression, thinking already dark areas would
get darker still. I just figured that would be how it would be, as though
illumination drops. Ultimately to zero. I had thought in terms of a light fade
for radiosity, of sorts. Like a shadow + dark color = darker shadow, or
something along those lines of thinking.

So this means the current way radiosity works is only to brighten upward from a
certain threshold (the actual pigment color, or lack of any) and never go lower?

I always feel like the kid in school class who was daydreaming and not paying
attention. Oh yeah, I was that kid.  LOL

Bob

Post a reply to this message

On 12/5/2016 12:33 PM, omniverse wrote:
> So this means the current way radiosity works is only to brighten upward from a
> certain threshold (the actual pigment color, or lack of any) and never go lower?


You could try negative light for darkening.

-- 

Regards
     Stephen

Post a reply to this message

Am 05.12.2016 um 13:33 schrieb omniverse:

> Sounds like I was under the wrong impression, thinking already dark areas would
> get darker still. I just figured that would be how it would be, as though
> illumination drops. Ultimately to zero. I had thought in terms of a light fade
> for radiosity, of sorts. Like a shadow + dark color = darker shadow, or
> something along those lines of thinking.
> 
> So this means the current way radiosity works is only to brighten upward from a
> certain threshold (the actual pigment color, or lack of any) and never go lower?

The lower threshold is actually the direct brightness from conventional
light sources. In full shadow, that would be zero.

But normally there's barely a crevice where radiosity-based illumination
doesn't find a way in.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Am 05.12.2016 um 13:33 schrieb omniverse:
>
> > Sounds like I was under the wrong impression, thinking already dark areas would
> > get darker still.
> >
> > So this means the current way radiosity works is only to brighten upward from a
> > certain threshold (the actual pigment color, or lack of any) and never go lower?
>
> The lower threshold is actually the direct brightness from conventional
> light sources. In full shadow, that would be zero.
>
> But normally there's barely a crevice where radiosity-based illumination
> doesn't find a way in.

Okay. Wrapping my head around this, I believe. I had been thinking in terms of a
dual effect, namely illumination and anti-illumination.

And as Stephen suggested, negative light color doesn't seem to adversely affect
radiosity, as in uncontrollable way, so maybe there are possibilities with that.

This had me trying negative emission and I was surprised by the result, although
perhaps I shouldn't be based on what you've been saying.

A test of 2 spheres, rgb 1 and rgb 0 both with emission -1, caused only a
darkened area for the white (now black) sphere on a white plane.
Black sphere with emission -1 made no change to the white plane below it.

I take from this that color > 0 is required for emission to do anything. This in
turn makes me think there should be some way to get a black or dark object to
enhance the darkness around it.

Post a reply to this message

On 12/5/2016 4:14 PM, omniverse wrote:
> And as Stephen suggested, negative light color doesn't seem to adversely affect
> radiosity, as in uncontrollable way, so maybe there are possibilities with that.

You can use a non reflecting diffuse thin box as a fill in like they do 
in photography. If you make its colour negative it sucks the light out. 
Or so it seemed to me when I tried it.

-- 

Regards
     Stephen

Post a reply to this message