Gilles Tran wrote in message <369F762C.CC65BE0A@inapg.inra.fr>...
>There would have been two solutions to implement volumetrics in povray : the
>first one (probably the one chosen in commercial packages) is to simulate
>them in a user-friendly but simplistic and limited way ; the second one
>(chosen by the pov-team) is to do it the hard way, by recreating real-life
>phenomena. This gives (advanced) povray users a complete control on media
>effects with one unique general syntax. Some media stuff we saw recently was
>really amazing so the choice is good news. The less good news is that media
>is obviously very difficult to master, and that simple effects are hard to
>achieve and require a lot trial-and-error, noticeably for those who don't
>understand the underlying physics (like me).
>


Maybe they should implement both methods, possibly making atmosphere a seperate thing
altogether.

Andy

Post a reply to this message

I've been planning on making something new to cover media, just haven't
found the time.  I'll tell you one thing though, it won't cover halos in
any shape or form.

I see it this way.  When it comes to scattering, media is just as good
as atmosphere.  They are both about equally bad in terms of speed, and
if you use jitter with atmosphere you have the same look.  So for at
least that aspect of media I'd say there's no major difference.

The other options in halos are a different story.  The emitting
and...well, just the emitting as the other two sucked...were actually
quite useful on account of the speed.  This old computer used to blow
right through an emitting halo in seconds and get smooth results.  Not
quite the same with media.  My understanding is that the sampling method
is different than with media.  Bringing back that method might be a good
idea as an option in media for things that don't need as much accuracy.  
So anyway, I got this thing I'm working on, plus that collection of
stars I promised, and this other thing...but I swear I'll try to cook
something up on media as soon as I can. :)

-Mike

Bob Hughes wrote:
> 
> Alas... I know poor Mike, through AOLs POV Chat, well. (Shakespearian,
> don't ya think?)
> 
> I've long since seen and re-seen those tutorial pages and it's
> unfortunate I can't make head nor tails of media even looking at other
> examples. It keeps defying my sense of logic. I'm still thinking it's a
> learning experience needed, halo and atmosphere in 3.0* was likewise, as
> is anything.
> From what I've been able to tell so far though 3.1 is not a perfect step
> up, as media can be much slower to produce an equal smoothness it seems.
> Well, not to be a whiner or anything, I just agree with many apparently
> that a quick and versatile method of using media in it's variety of ways
> has become a holy grail to seek out.
> Color manipulation of media is, itself, like a whole new world.
>

Post a reply to this message

Please don't think I'm pressuring you. All I was getting at is, so far
I've not increased *my* knowledge level much with or without looking
over those tuts. But then, it seems others have perhaps heaped a burden
on you to continue your tutorial pages, so by all means I'm not standing
in your way either, hint-hint.


Mike wrote:
> 
> I've been planning on making something new to cover media, just haven't
> found the time.  I'll tell you one thing though, it won't cover halos in
> any shape or form.
> 
> I see it this way.  When it comes to scattering, media is just as good
> as atmosphere.  They are both about equally bad in terms of speed, and
> if you use jitter with atmosphere you have the same look.  So for at
> least that aspect of media I'd say there's no major difference.
> 
> The other options in halos are a different story.  The emitting
> and...well, just the emitting as the other two sucked...were actually
> quite useful on account of the speed.  This old computer used to blow
> right through an emitting halo in seconds and get smooth results.  Not
> quite the same with media.  My understanding is that the sampling method
> is different than with media.  Bringing back that method might be a good
> idea as an option in media for things that don't need as much accuracy.
> So anyway, I got this thing I'm working on, plus that collection of
> stars I promised, and this other thing...but I swear I'll try to cook
> something up on media as soon as I can. :)
> 
> -Mike
> 
> Bob Hughes wrote:
> >
> > Alas... I know poor Mike, through AOLs POV Chat, well. (Shakespearian,
> > don't ya think?)
> >
> > I've long since seen and re-seen those tutorial pages and it's
> > unfortunate I can't make head nor tails of media even looking at other
> > examples. It keeps defying my sense of logic. I'm still thinking it's a
> > learning experience needed, halo and atmosphere in 3.0* was likewise, as
> > is anything.
> > From what I've been able to tell so far though 3.1 is not a perfect step
> > up, as media can be much slower to produce an equal smoothness it seems.
> > Well, not to be a whiner or anything, I just agree with many apparently
> > that a quick and versatile method of using media in it's variety of ways
> > has become a holy grail to seek out.
> > Color manipulation of media is, itself, like a whole new world.
> >

-- 
 omniVERSE: beyond the universe
  http://members.aol.com/inversez/POVring.htm
=Bob

Post a reply to this message

Bob Hughes wrote in message <369ECBDE.8FD96178@aol.com>...
>Anyone get a reasonable, and useable, atmospheric media working for
>various situations? I.e., different camera/object distances and number
>of lights so when you make a new scene the media (as used for
>atmosphere) doesn't wreak havoc while adjusting it endlessly?
>I understand the basic scaling syndrome of keeping media density in sync
>with container objects, but I've been lost too many times trying the
>containerless "air", since each scene is not the same as the last.


One thing you have to remember is to scale the densities according to
the scale of the scene. I've found that to have 50% of light absorbed
by a media of thickness "dist" the appropriate definition is
  media { absorption 1/(sqrt(2)*dist) }
I suppose the same applies to scattering media. Since the density is
constant, you can speed up rendering by using "intervals 1". Oh, and
don't use the fade_distance keyword in interior if you want a realistic
scene - the formula used has nothing to do with realistic absorption.

I do agree that the overlap between the values in density{} and the
actual absorption/emission/scattering values can be a bit confusing.
It may be a good idea to treat either of the values as being normalized
to the 0.0-1.0 range and to vary the other.

BTW, whatever were the povray designers thinking when they implemented
light_source fading? For some reasong the realistic formula
  attenuation = pow( fade_distance/d, fade_power )
was replaced by
  attenuation = 2/( 1 + pow( d/fade_distance, fade_power ) )
which has AFAIK absolutely no connection to reality, even though it does
approximate the proper formula at distances far greater than fade_distance
as long as you remember to divide the brightness by two.. (The same cannot
be said for the aforementioned interior fading.)

Hmm.. maybe I should finally get a C compiler for this computer and start
writing my own patches.. my wish list is growing too long to remember.

--
Ilmari Karonen (ilt### [at] scifi)
(not r### [at] hakasaunalahtifi as the IRTC comments list claims..)
http://www.sci.fi/~iltzu/

Post a reply to this message

Ilmari Karonen wrote:
> 
> Bob Hughes wrote in message <369ECBDE.8FD96178@aol.com>...
> >I understand the basic scaling syndrome of keeping media density in sync
> >with container objects, but I've been lost too many times trying the
> >containerless "air", since each scene is not the same as the last.
> 
> One thing you have to remember is to scale the densities according to
> the scale of the scene. I've found that to have 50% of light absorbed
> by a media of thickness "dist" the appropriate definition is
>   media { absorption 1/(sqrt(2)*dist) }

Okay, I'll try that out. Thanks for keeping me going. (I did know to
scale, maybe everyone will catch on to that procedure eventually)

> I suppose the same applies to scattering media. Since the density is
> constant, you can speed up rendering by using "intervals 1". Oh, and
> don't use the fade_distance keyword in interior if you want a realistic
> scene - the formula used has nothing to do with realistic absorption.
> 

Watch out with the intervals being lower than the number of
light_sources. Not sure yet of the correlation, but I have had crashes
using a intervals integer less than the lights number.

> BTW, whatever were the povray designers thinking when they implemented
> light_source fading? For some reasong the realistic formula
>   attenuation = pow( fade_distance/d, fade_power )
> was replaced by
>   attenuation = 2/( 1 + pow( d/fade_distance, fade_power ) )
> which has AFAIK absolutely no connection to reality, even though it does
> approximate the proper formula at distances far greater than fade_distance
> as long as you remember to divide the brightness by two.. (The same cannot
> be said for the aforementioned interior fading.)
> 

Any POV Team member know why or can confirm this?

> --
> Ilmari Karonen (ilt### [at] scifi)
> (not r### [at] hakasaunalahtifi as the IRTC comments list claims..)
> http://www.sci.fi/~iltzu/

-- 
 omniVERSE: beyond the universe
  http://members.aol.com/inversez/POVring.htm
=Bob

Post a reply to this message

Gilles Tran wrote:

> The less good news is that media is obviously very difficult to master, and
that simple
> effects are hard to achieve and require a lot trial-and-error, noticeably for
those who
> don't understand the underlying physics (like me).

I sincerely applaud the POV Team for making a more difficult, real-life
phenomena ray tracer.

> One solution would be for one of our media wizards to build a library of
> specialised macros for particular situations.

No! If I wanted to lose that much respect for myself, I would spend my next
bonus on 3DS and just push buttons to render great art. I'd rather that half of
the amount of effort that Gilles just suggested be put into a comprehensive
tutorial on media. I like to figure out hard things, especially related to my
engineering background.   I also like to know what I can expect to be able to
do.

What is media? The POV 3.1 documentation took a huge step backwards compared to
3.02.

Post a reply to this message