Creating realistic natural objects like trees. 

	I've seen a handful of interesting to "how'd he do that" results. 

	However, I think everyone agrees they all lack something including the
developers. 

	Take for example a symetrical christmas tree, perfectly symetrical
being the search on December 20th. But there is a good and bad side.
Also a real tree has good and bad seasons over its years of growing. A
real tree tries to maximize its sunlight. A real tree has other tree
competing for sunlight. 

	So could the "growth" algorithms for trees be modified for nearby trees
while small and taking over when tall, shading others? Prevailing winds
to minimize growth on one side? Dry summers resulting in lesser growth? 

	I am not suggesting an [tree expert]-oligist simulation (if you can,
publish professionally) but rather not a simple random factor but
natural random factors. Two to three good and bad growing seasons tend
to come in a row so branch spacing should vary with that in mind instead
of purely randomly. Growth on one side should be less than on another.
There should be signs of early competition for sunlight. 

	Those should not be hard to weight just by playing around given some of
the algorithms I have looked at. Consideration of real life growth
conditions will be more difficult but there should be a simple set of
variations that mimic what really happens when as trees grow that can
make them look a bit more realistic. 

	One side comment, "forests" of the same kind of tree are certainly man
made. 

-- 
None of my opinions are humble. 
	-- The Iron Webmaster, 96

Post a reply to this message

I would disagree with the comment about looking for a perfectly symmetrical
Christmas tree. That is exactly what I don't look for, because I put up the
Christmas tree in a corner in a fairly small room, so I need a tree where
approximately a quarter of the circumference has lower branches that are either
shorter or more upright than the other three quarters.

Matt Giwer (almost) wrote:

>
>         Take for example a symmetrical Christmas tree, perfectly symmetrical
> being the search on December 20th. But there is a good and bad side.
> Also a real tree has good and bad seasons over its years of growing. A
> real tree tries to maximize its sunlight. A real tree has other trees
> competing for sunlight.
>

Post a reply to this message

Richard Morton wrote:
> 
> I would disagree with the comment about looking for a perfectly symmetrical
> Christmas tree. That is exactly what I don't look for, because I put up the
> Christmas tree in a corner in a fairly small room, so I need a tree where
> approximately a quarter of the circumference has lower branches that are either
> shorter or more upright than the other three quarters.
> 
> Matt Giwer (almost) wrote:

> >         Take for example a symmetrical Christmas tree, perfectly symmetrical
> > being the search on December 20th. But there is a good and bad side.
> > Also a real tree has good and bad seasons over its years of growing. A
> > real tree tries to maximize its sunlight. A real tree has other trees
> > competing for sunlight.

	An oldie but a goodie. 

	I have been known to ramble on with examples of what lead me to an idea
rather than what follows from it.

	I think that one is in reference to realistic trees for rendering. My
point was that a perfectly symetrical tree is unnatural as is random
asymetry. For example, the sun is always on one side of a christmas
tree, there will be prevailing winds and outside of tree farms growth
will be hindered where it is in the shadows of other trees. 

	The bottom line is that simulated trees should be asymetrical in a
manner emulating the reasons they are asymetrical in nature in order to
look realistic. 

	I don't know how to do that. I was just commenting one potential reason
for the lack of realism of the best models. 

-- 
Just what is the difference between anti-cathoic 
and anti-semitic?
	-- The Iron Webmaster, 298

Post a reply to this message