"clipka" <nomail@nomail> wrote in message 
news:web.49d5dcd34ee6dd4ef708085d0@news.povray.org...
> "Zeger Knaepen" <zeg### [at] povplacecom> wrote:
>> how about making the default value for emission the same as the
>> ambient-value?
>
> No, definitely not. In a radiosity scene, only very, very few textures are
> typically intended to emit light. But in a non-radiosity scene, you 
> probably
> want many, many textures to have an ambient term, in order to approximate
> "ambient illumination" (i.e. illumination by light scattered diffusely 
> from
> other objects).
>
> So the typical use case would be to have ambient X emission 0.

this would completely break older scenes.  If the default emission-value is 
the same as the specified ambient-value, then older scenes will render 
exactly the same as they used to.

>> Also, I don't think emission should have any meaning when not used with
>> radiosity.
>
> I disagree.
>
> In a radiosity-only scene, of course you want emission to have an effect,
> because there'd be no other way to get light into the scene (except for a 
> sky
> sphere).
>
> When lighting the same scene classically, you probably want the same thing 
> to
> look similar when directly visible in the scene. For this, it will have to
> emit, too.

They will look exactly the same when directly visible, since only ambient 
(and not emission) would affect the look of the texture itself.
So finish {ambient 1 emission .5} and finish {ambient 1 emission 5} will 
look the same in a non-radiosity scene, but will affect their environment 
different in a radiosity-scene... but even there the texture itself will 
look exactly the same

> Christian's idea to leave ambient fully functional in radiosity scenes for
> compatibility, and expecting the user to actively turn it off by setting
> ambient_light to 0, seems the most viable solution to me.

as already stated, that wouldn't work in a scene without convention 
light_sources

I'm all for an emission-value in the finish-statement, but only if it
a) doesn't just duplicate the effect of ambient (it has to have a different 
meaning than ambient) and
b) doesn't break old scenes (older scenes have to render exactly the same as 
they do now)
and the only way, IMHO, to satisfy both conditions, is to let emission (and 
only emission) only have effect in radiosity-lighting and to let the default 
emission-value be the same as the specified ambient-value.

cu!
-- 
#macro G(b,e)b+(e-b)*C/50#end#macro _(b,e,k,l)#local C=0;#while(C<50)
sphere{G(b,e)+3*z.1pigment{rgb G(k,l)}finish{ambient 1}}#local C=C+1;
#end#end _(y-x,y,x,x+y)_(y,-x-y,x+y,y)_(-x-y,-y,y,y+z)_(-y,y,y+z,x+y)
_(0x+y.5+y/2x)_(0x-y.5+y/2x)            // ZK http://www.povplace.com

Post a reply to this message

"Zeger Knaepen" <zeg### [at] povplacecom> wrote:
> > So the typical use case would be to have ambient X emission 0.
>
> this would completely break older scenes.  If the default emission-value is
> the same as the specified ambient-value, then older scenes will render
> exactly the same as they used to.

I don't see how setting emission to 0 would break older scenes. Not with
Christian's suggestion for keeping backward-compatibility.

To recap the combined suggestion:

- Ambient works as it has always done. It's just deprecated to be used for
anything that is supposed to look like it's actively emitting light (as opposed
to things that are supposed to look like they're passively illuminated by
indirect light).

- For things that are supposed to look like they're actively emitting light, it
will be recommended to use the new "emission" keyword instead, which does
exactly the same as ambient except that it is not affected by the ambient_light
setting. (If both are used, the effects add up.)

- In non-radiosity scenes, nothing else changes to the current situation.

- In radiosity-only scens, it will be recommended to set ambient_light to 0; any
other value will only be accepted for backward compatibility, and generate a
deprecation warning. It will be expected that the emission mechanism is used to
generate light sources, instead of the ambient mechanism.

- In combined scenes, too, it will be recommended to set ambient_light to 0, and
any other value generate a deprecation warning. For such scenes, it will be
expected that the emission mechanism without a classic light source is used to
generate visible light sources, but otherwise classic light sources are used.

For this approach to work, clearly emission must default to 0.

> > When lighting the same scene classically, you probably want the same thing
> > to
> > look similar when directly visible in the scene. For this, it will have to
> > emit, too.
>
> They will look exactly the same when directly visible, since only ambient
> (and not emission) would affect the look of the texture itself.

Ah, so this is where you're misunderstanding things (or have different
expectations): You expect "emission" to only affect radiosity, and not the way
the thing looks when directly visible in the shot.

I can tell you that (a) I wouldn't want it that way (because after all, the
directly visible glow of an object *is* emission, not ambient illumination, so
it would be right to control both with the "emission" statement, and (b) it
would also be more difficult to implement, because without additional
programming effort, radiosity generally "sees" what the observer sees as well.


So both ambient *and* emission would have a directly visible effect, and both
would affect radiosity. It's just that in radiosity scenes the one would be
suggested to be turned off, while the other would remain effective.


Now you may wonder, "but what if I want my texture to have that slight ambient
term in radiosity scenes, too?"

Well, you just wouldn't. Ambient illumination as modeled with the "ambient" and
"ambient_light" mechanism is just an approximation of what radiosity does
automatically and better; with radiosity, that mechanism is simply just in the
way. You don't need it to model realistic textures for radiosity scenes. To the
contrary: You *can't* model realistic (non-glowing) textures for radiosity
scenes that have any ambient term.


> > Christian's idea to leave ambient fully functional in radiosity scenes for
> > compatibility, and expecting the user to actively turn it off by setting
> > ambient_light to 0, seems the most viable solution to me.
>
> as already stated, that wouldn't work in a scene without convention
> light_sources

That would depend on the approach used; with the approach Christian and I have
in mind, it would work perfectly.


> I'm all for an emission-value in the finish-statement, but only if it
> a) doesn't just duplicate the effect of ambient (it has to have a different
> meaning than ambient) and

It will - if only due to the fact that it will be independent of the
ambient_light, which can then be used to switch off all ambient terms globally,
while leaving a mechanism to model glowing objects.

And (likewise important if I'm asked), despite all similarities in effect, it
will have a totally different *meaning* - thank you for the word.

The "ambient" term will state - in line with its original intention - that an
object receives ambient illumination from somewhere, and needs therefore to be
non-black even in shadows. The very same thing radiosity computes automatically
and at much better quality.

The "emission" term will state that an object emits light by itself regardless
of other illumination sources; "ambient" has been mis-used for this purpose in
the past in lack of a separate mechanism, but this is the very reason why
ambient has become a problem.


> b) doesn't break old scenes (older scenes have to render exactly the same as
> they do now)

Doesn't happen with my idea, when combined with Christian's suggestion.


> and the only way, IMHO, to satisfy both conditions, is to let emission (and
> only emission) only have effect in radiosity-lighting and to let the default
> emission-value be the same as the specified ambient-value.

IYHO. I disagree, for multiple reasons:

(1) Semantics: Passive ambient illumination *is* something different in the real
world than active emission of light; while in the approach you have in mind,
neither "ambient" nor "emission" would properly describe what the things do
(because neither would have any real-world equivalent, taken on its own).

(2) Ease of use: With the approach proposed by Christian and me, the bulk of
existing textures would be immediately radiosity-ready. No need to change
anything, unless the texture is *intended* to glow - just set ambient_light to
0 and render; and making full use of that approach in future scenes would be
virtually no more difficult than it is now: Just throw in an "emission"
statement for every texture *intended* to glow, and do the others as usual.

With your approach, the bulk of existing textures would have to be explicitly
equipped with "emission 0" statements to make them radiosity-ready - and future
scenes would of course require that statement, too, to make the best use of the
approach. Worse yet: A significant portion of textures will continue to be
created non-radiosity-ready, because scene designers typically not using
radiosity will often just not bother adding that "emission 0" statement.

(3) Ease of implementation: Any surface feature that should "look" different to
radiosity than it does to the observer requires extra implementation effort.


To sum it up: I don't think the approach you had in mind is really viable.

Post a reply to this message

I almost completely disagree with your post, but I'm to tired for a 
discussion now :)

cu!
-- 
#macro G(b,e)b+(e-b)*C/50#end#macro _(b,e,k,l)#local C=0;#while(C<50)
sphere{G(b,e)+3*z.1pigment{rgb G(k,l)}finish{ambient 1}}#local C=C+1;
#end#end _(y-x,y,x,x+y)_(y,-x-y,x+y,y)_(-x-y,-y,y,y+z)_(-y,y,y+z,x+y)
_(0x+y.5+y/2x)_(0x-y.5+y/2x)            // ZK http://www.povplace.com

Post a reply to this message

clipka nous illumina en ce 2009-04-03 05:54 -->
> "Zeger Knaepen" <zeg### [at] povplacecom> wrote:
>> how about making the default value for emission the same as the
>> ambient-value?
> 
> No, definitely not. In a radiosity scene, only very, very few textures are
> typically intended to emit light. But in a non-radiosity scene, you probably
> want many, many textures to have an ambient term, in order to approximate
> "ambient illumination" (i.e. illumination by light scattered diffusely from
> other objects).
> 
> So the typical use case would be to have ambient X emission 0.
> 
> 
>> Also, I don't think emission should have any meaning when not used with
>> radiosity.
> 
> I disagree.
> 
> In a radiosity-only scene, of course you want emission to have an effect,
> because there'd be no other way to get light into the scene (except for a sky
> sphere).
> 
> When lighting the same scene classically, you probably want the same thing to
> look similar when directly visible in the scene. For this, it will have to
> emit, too.
> 
> 
> Christian's idea to leave ambient fully functional in radiosity scenes for
> compatibility, and expecting the user to actively turn it off by setting
> ambient_light to 0, seems the most viable solution to me.
> 
> 
Setting ambient_lights 0 is definetly NOT a viable solution. As you said, it 
effectively turn OFF any and all ambient, for every finish, even those you WANT 
to have an ambient value and emit light.

-- 
Alain
-------------------------------------------------
   I knew a girl so ugly that she was known as a two-bagger. That's When you put 
a bag over your head in case the bag over her head comes Off.
	Rodney Dangerfield

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> To recap the combined suggestion:
>
> [snip]

Even though my novice vote should not count for much, I just want to
chip in to say that I am in favour of clipka's and Cristian's combined
suggestion. I agree that the meaning of the ambient and emmision statement is
clearly distinct, and both meanings are easy to comprehend. Also, I cannot see
anything awkward in using old textures, defining textures that work in normal
rendered scenes as well as radiosity-rendered scenes, re-rendering old scenes,
etc when this approach would be implemented.

Cheers,
Erwin

Post a reply to this message

Alain <ele### [at] netscapenet> wrote:
> > Christian's idea to leave ambient fully functional in radiosity scenes for
> > compatibility, and expecting the user to actively turn it off by setting
> > ambient_light to 0, seems the most viable solution to me.
> >
> Setting ambient_lights 0 is definetly NOT a viable solution. As you said, it
> effectively turn OFF any and all ambient, for every finish, even those you WANT
> to have an ambient value and emit light.

.... which is *exactly* what we'd have the "emission" statement for, so no
problem.

Post a reply to this message

"clipka" <nomail@nomail> wrote:
> .................

Really an interesting approach, clipka.
But *emission* keyword can be confusing to the *emission* media?
(only in terms of keywords, obviously)

--
Carlo

Post a reply to this message

"Carlo C." <nomail@nomail> wrote:
> Really an interesting approach, clipka.
> But *emission* keyword can be confusing to the *emission* media?
> (only in terms of keywords, obviously)

Indeed; so far, I considered the keyword only a kind of "working draft", and I
wasn't sure about it either.

On the other hand, using this particular keyword may actually be of benefit: You
put "emission COLOR" in a media, and the object's interior glows; you put
"emission COLOR" in a surface, and that surface glows. Sounds reasonably
straightforward to me.

Post a reply to this message

Carlo C. nous illumina en ce 2009-04-04 03:39 -->
> "clipka" <nomail@nomail> wrote:
>> .................
> 
> Really an interesting approach, clipka.
> But *emission* keyword can be confusing to the *emission* media?
> (only in terms of keywords, obviously)
> 
> --
> Carlo
> 
> 
If you have a media with emission in a radiosity scene with media on, that media 
will emit light that will illuminate it's surrounding.

So, the proposed use of emission in a finish have the same meaning and purpose.

-- 
Alain
-------------------------------------------------
You know you've been raytracing too long when you see something in the real 
world and you think, "Hey! How did they get that effect?"

Post a reply to this message

Alain <ele### [at] netscapenet> wrote:
> >
> >
> If you have a media with emission in a radiosity scene with media on, that media
> will emit light that will illuminate it's surrounding.
>
> So, the proposed use of emission in a finish have the same meaning and purpose.
>
> --
> Alain
> -------------------------------------------------
> You know you've been raytracing too long when you see something in the real
> world and you think, "Hey! How did they get that effect?"


True, now I have things more clear. ;-)
Thanks Alain, and thanks Clipka.

--
Carlo

Post a reply to this message