Someone said that the output_alpha=on is supposed to render the scene so
the background color doesn't influence the scene. The following scene
looks much different if you set the background to, say, rgb<0,1,0>.
Apparently the background affects the color of reflected rays (although
experimentation by me seems to show that backgrond color does not affect
refracted rays or media samples, unless of course there's reflection on
the same ray.) I haven't tried radiosity with +ua yet.

btw, I'm working on a win2k machine, x86, 128MB RAM

John Haiducek

sphere{
  0,1
  pigment{rgbt <.8,.9,1,1>}
  finish{reflection .25}
  hollow
}

camera{
  location <0,0,-3>
  look_at 0
}

light_source{
  <-20,40,-20>
  rgb 1
}

background{rgb <0,1,0>}

Post a reply to this message

John David Haiducek <jha### [at] enginumichedu> wrote:
: Someone said that the output_alpha=on is supposed to render the scene so
: the background color doesn't influence the scene.

  Not really. What happens with +ua is that the background color has no
effect on antialiased pixels as well as (semi)transparent objects.
  There's no reasonable way to use the alpha channel with reflections. A
reflective opaque object is not transparent and thus it should not have any
alpha value. If it reflectes the "background", then what color should be
the reflection? The color of the background, of course.

  This is the correct behaviour and there's nothing wrong in it.

-- 
#macro N(D,I)#if(I<6)cylinder{M()#local D[I]=div(D[I],104);M().5,2pigment{
rgb M()}}N(D,(D[I]>99?I:I+1))#end#end#macro M()<mod(D[I],13)-6,mod(div(D[I
],13),8)-3,10>#end blob{N(array[6]{11117333955,
7382340,3358,3900569407,970,4254934330},0)}//                     - Warp -

Post a reply to this message

I can think of some situations where using the alpha channel for
reflection could be useful. (If you consider this a feature request, I
apologize.) For instance, in case one wanted to use a reflection to create
a distorted image of an object. Perhaps the renderer could use the
transmit component of the background color to determine how reflection is
handled (i.e. if the transmittance was 1, completely
reflective objects would be transparent.)

Since I didn't say so before, I'm really excited about the improved alpha
output. It's a powerful feature in any case.

John Haiducek

Post a reply to this message

I think the reason reflection isn't transparent is because if it were, and
you had the PNG on, say, a web page, the object would appear transparent
rather than reflective. It would show things behind it when it should show
reflected things.

- Slime
[ http://www.slimeland.com/ ]
[ http://www.slimeland.com/images/ ]

"John David Haiducek" <jha### [at] enginumichedu> wrote in message
news:Pin### [at] blueenginumichedu...
> I can think of some situations where using the alpha channel for
> reflection could be useful. (If you consider this a feature request, I
> apologize.) For instance, in case one wanted to use a reflection to create
> a distorted image of an object. Perhaps the renderer could use the
> transmit component of the background color to determine how reflection is
> handled (i.e. if the transmittance was 1, completely
> reflective objects would be transparent.)
>
> Since I didn't say so before, I'm really excited about the improved alpha
> output. It's a powerful feature in any case.
>
> John Haiducek
>

Post a reply to this message

Slime <noo### [at] hotmailcom> wrote:
: I think the reason reflection isn't transparent is because if it were, and
: you had the PNG on, say, a web page, the object would appear transparent
: rather than reflective. It would show things behind it when it should show
: reflected things.

  Yes, that's exactly why alpha in reflections doesn't make any sense.

  Also I like a lot the improved alpha channel support. Now I can create
media flames with transparency to a PNG at it will work like a marvel.
  For example I made this alpha channel test page using exactly that:

http://www.cs.tut.fi/~warp/alphatest.html

  If your browser supports transparent PNGs correctly, the page should look
like this:

http://www.cs.tut.fi/~warp/alphatestsnapshot.jpg

  So far I know of just few browsers which have this good alpha channel
support: Netscape 6.x, Mozilla 0.9.x, Internet Explorer 5.x for Macintosh
and perhaps a couple of less-known browsers.
  Note that IE 5.x for Windows does *not* support PNG alpha channel correctly.

-- 
#macro N(D,I)#if(I<6)cylinder{M()#local D[I]=div(D[I],104);M().5,2pigment{
rgb M()}}N(D,(D[I]>99?I:I+1))#end#end#macro M()<mod(D[I],13)-6,mod(div(D[I
],13),8)-3,10>#end blob{N(array[6]{11117333955,
7382340,3358,3900569407,970,4254934330},0)}//                     - Warp -

Post a reply to this message

Thaks. I agree now that there isn't a logical way to use alpha on
reflected rays. It makes even less sense when there are objects behind
the reflective object.

John Haiducek

Post a reply to this message

Warp wrote in message <3be3e176@news.povray.org>...
>  So far I know of just few browsers which have this good alpha channel
>support: Netscape 6.x, Mozilla 0.9.x, Internet Explorer 5.x for Macintosh
>and perhaps a couple of less-known browsers.


Opera 5.11 does NOT support PNG alpha channel.

--
Mark

Post a reply to this message