Andrew Cocker wrote:
> 
> I just had an idea. See what you think. There has been much discussion in this group
> regarding media and it's shortcomings, particularly with regard to creating a dusty
/
> smoky atmosphere. This is what I use it for mainly, but usually give up because a)
it
> takes so long to render, and b) because it often / usually looks too grainy.
> My idea is this. Would it not be effective to use similar code to the way that
constant
> fog works ( which seems to just change each pixel by a certain amount towards the
chosen
> fog colour ), but *only for those pixels that are in line of sight of a light
source*. So
> any obstructions would effectively seem to be casting a shadow *onto* the fog. It
seems to
> me that this is how a dusty atmosphere looks when light shines in through a window
etc. Of
> course the huge benefit of this approach would be a minimal increase in rendering
time.
> You could have a parameter for controlling the intensity of the effect ie how near
to
> white each pixel becomes, a parameter to control how the effect fades as it moves
further
> from the light source. I'm not sure how additional lights would affect things..they
could
> each add to the intensity where spotlights cross one another, or they could all be
> automatically adjusted so an overall effect is maintained, regardless of the number
of
> lights used.
> Finally, the fog could be affected by, for example, a pigment map in a similar way
to a
> density map, so that wherever rgbf 1 occured, the fog was clear. This could then be
> animated in the usual way to create many atmospheric conditions. Sort of a simpler
> Density.
> 
> Have I overlooked something here? My idea seems so simple, but sometimes the simple
ideas
> work best. I'm not suggesting getting rid of media, as it excels at fires, clouds
etc, but
> in my opinion it's overkill in some instances.
> 
> Maybe for the next version?
> 
> -----------
> Andy

  This topic and the solutions you have suggested were discussed in
some detail back in January. I believe the thrust of the discussion
centered aorund the use of monte carlo sampling and it's limitions.
I tmight have been Nathan Kopp who mentioned tht he has been trying
different sampling methods to be used with the media feature and has
passed on the concerns of people about the process to Chris Young
the pov team coordinator. I guess you could say there is work in
progress to add more functinality to the media process but I have
no confirming data to support this. I do agree that there are some
sampling methods like that used in the previous atmosphere feature that
could add greater funtionality to the current system and it would be
nice to see some improvements in that area.

-- 
Ken Tyler

mailto://tylereng@pacbell.net

Post a reply to this message