Is there some combination of no_image no_reflection tags that will allow an
object to reflect onto one scene element but not onto another? Or some other
method? Like in this sample picture (two images combined) where the blue ball
reflects onto the silver ball but not the red ball. Maybe it would be cool if
the concept of the light_group was extended to a "scene_group" where only
objects in the group reacted with each other regarding reflections, radiosity,
shadows, etc. That might be useful for special effects.

Regards,
Dave Blandston

Post a reply to this message

Attachments:
Download 'temp.png' (261 KB)

Preview of image 'temp.png'

"Dave Blandston" <nomail@nomail> wrote in message 
news:web.4cf3d937758d7bd8966554e20@news.povray.org...
> Is there some combination of no_image no_reflection tags that will allow 
> an
> object to reflect onto one scene element but not onto another? Or some 
> other
> method? Like in this sample picture (two images combined) where the blue 
> ball
> reflects onto the silver ball but not the red ball. Maybe it would be cool 
> if
> the concept of the light_group was extended to a "scene_group" where only
> objects in the group reacted with each other regarding reflections, 
> radiosity,
> shadows, etc. That might be useful for special effects.
>
> Regards,
> Dave Blandston
>

Ha!  I've been wanting a "scene_group" type group for ages, it would be 
awesome to be able to combine stuff into groups like that and not effect 
other groups settings etc...

Cheers Andreas

Post a reply to this message

On 11/29/2010 11:47 AM, Dave Blandston wrote:
> Maybe it would be cool if
> the concept of the light_group was extended to a "scene_group" where only
> objects in the group reacted with each other regarding reflections, radiosity,
> shadows, etc. That might be useful for special effects.

It could also be useful for making scenes render quickly.

Regards,
John

Post a reply to this message

Dave Blandston wrote:

> Or some other method?

I wonder if the aoi pattern should be extended
to support an absolute direction vector which is
then compared to the incoming ray (ignoring the
surface normal). Then the complete appearance of
the sphere can change depending on the viewpoint
(in particular it could become invisible).

The current aoi pattern is not very helpful here
as every viewpoint sees points on the sphere at all
angles from 0 to 90 degrees.

Post a reply to this message

Le 01/12/2010 00:57, Christian Froeschlin a écrit :
> Dave Blandston wrote:
> 
>> Or some other method?
> 
> I wonder if the aoi pattern should be extended
> to support an absolute direction vector which is
> then compared to the incoming ray (ignoring the
> surface normal). Then the complete appearance of
> the sphere can change depending on the viewpoint
> (in particular it could become invisible).
> 
> The current aoi pattern is not very helpful here
> as every viewpoint sees points on the sphere at all
> angles from 0 to 90 degrees.


I will think more about it, but no promise.
May be it would then need a better name than "angle of incidence" (aoi)
as it would then ignore the actual surface. any suggestion ?

would expect another range than 0 to 90 degrees ?

Post a reply to this message

Le_Forgeron wrote:

> May be it would then need a better name than "angle of incidence" (aoi)
> as it would then ignore the actual surface.

Maybe the keyword aoi could be kept but in the documentation
described as "angle of incident ray", which it still is, just
not necessarily in relation to the surface normal.

> would expect another range than 0 to 90 degrees ?

0 to 180 would seem appropriate.

Post a reply to this message

> Dave Blandston wrote:
>
>> Or some other method?
>
> I wonder if the aoi pattern should be extended
> to support an absolute direction vector which is
> then compared to the incoming ray (ignoring the
> surface normal). Then the complete appearance of
> the sphere can change depending on the viewpoint
> (in particular it could become invisible).
>
> The current aoi pattern is not very helpful here
> as every viewpoint sees points on the sphere at all
> angles from 0 to 90 degrees.

In a way, the slope pattern does take a vector to determine an absolute 
direction against whitch the normal is evaluated. That vector can have 
any orientation. That coupled with an inside_texture that is totaly 
transparent...

Then, it should be possible to have a slope that contains an aoi pattern 
in a pattern_map.

It gets realy touchy if an object is to be invisible from more than one 
other reflective objects and those are not close to each other.



Alain

Post a reply to this message

Le 01/12/2010 23:32, Christian Froeschlin nous fit lire :
> Le_Forgeron wrote:
> 
>> May be it would then need a better name than "angle of incidence" (aoi)
>> as it would then ignore the actual surface.
> 
> Maybe the keyword aoi could be kept but in the documentation
> described as "angle of incident ray", which it still is, just
> not necessarily in relation to the surface normal.

creating confusion with the actual aoi (actual = 3.7), not a good idea.

There is an issue inherent with such pattern: ulterior transformation
cannot affect the reference direction. which means that rotating an
object actually changes its appearance. A blue (as determined by the
color map) face rotated by 30° would be red (... or whatever).
It might be rather confusing when cloning objects around: the shapes are
cloned, but the textures appears to be different when more than a
translation+scale is applied (rotation & shear)!

With such behaviour, a name like "compass", "gyrocompass",
"astrocompass", or wiking proverbial "sun-stone" are more matching than
aoi. But they are a bit long (and sun-stone has an issue due to the - )
(iolite & labradorite)

polychroism (not pleochroism, as pleochroism need polarised light) might
be a name too. (but it would better describe aoi)

> 
>> would expect another range than 0 to 90 degrees ?
> 
> 0 to 180 would seem appropriate.
> 
>

Post a reply to this message

Le 02/12/2010 08:46, Le_Forgeron nous fit lire :
> Le 01/12/2010 23:32, Christian Froeschlin nous fit lire :
>> Le_Forgeron wrote:

>>> would expect another range than 0 to 90 degrees ?
>>
>> 0 to 180 would seem appropriate.

aov ? (angle of view)

Post a reply to this message

Le_Forgeron wrote:

> There is an issue inherent with such pattern: ulterior transformation
> cannot affect the reference direction. which means that rotating an
> object actually changes its appearance.

Yes, it's the same with the slope pattern. Also, the slope pattern
supports both vector and point source in 3.7, so an alternate vector
mode for aoi would not be completely unexpected

   aoi {relative_to <vector>}?

But I have no objections to "aov" as well.

Post a reply to this message