Tips for creation of 3D plants :
Before I prepare some real tutorials for this section, a couple of tips
that may save you a lot of nerves :
Don't overcomplicate the graph trying to add too much functionality into
It is better to create 5 separate variations of a plant, than have them
all inside one model.
TPF provides you with external parameters like age, health, season. So
in theory you should be able to create a model which generates trees from
30 cm tall seedling, through 20m lush adult to dead tree with sparse bald
broken branches. Yes, it is possible, but it does not mean it should be
done with every model.
Let's consider analogy to cars - One could create a model of a car that
could transform itself from a small compact to a huge truck, and then to
a road wreck. But surprisingly nobody does it. Simply because it does not
make much sense. It is better, simpler, easier to create 3 separate models
than one complicated "transformer" like that.
There may be some rare cases when such age/season/health functionality
makes sense, then do it if you can't go the other, simple way (create several
sets of plants)
You may try to implement one parameter like age, and if graph is still
manageable, try to add another. But think if next year you will be still
capable to figure out how it works.
With procedural models there is intrinsic risk of data loss.
Minute modification to software code has potential to destroy hundreds models.
Imagine that branch angle 0 meant branch orthogonal to parent. But for next
version of software, some developer gets an idea of changing this definition,
and now angle 0 means branch parallel to parent....
If you have just 3 models, each with a couple of nodes without complex,
extracted functions, maybe you can remade them in 15 minutes. But imagine
that over several years you have created a collection of 500 models, each
model is a complex maze of multiple interacting nodes, and after you updated
your software, all models look like.... Disaster ! When incompatibility
strikes, it may happen that remaking the plant from scratch is faster than
finding the cause of malfunction. This is why I suggest that after you create
the final model, you bake it to polygons and save in some common format
like OBS or 3DS. If incompatibility happens, at least you will keep the
static polygonal mesh.
TPF has some provisions to cope with backward compatibility when nodes
code is modified, but the risk still exists.
Check links to external tutorial in LINKS / TUTORIALS section. There is
already plenty of TPF tutorials prepared by Vladimir Chopine, and also quite
a lot of Speed Tree tutorials. So I don't know if I'll do mine TPF tutorials
- it would just double what Vladimir already did - pointless. Probably e-on
is also about to prepare some - there is already one on e-on youtube channel.
Maybe in this section I'll limit myself to write down some interesting
observations, suggestions helpful for TPF users.