[fe] Something like SpeedTree?

[Jason Weber]

On Sun, Mar 29, 2009 at 11:16 AM, Juhana Sadeharju
<Juhana.Sadeharju at uta.fi> wrote:
>
>
> On Fri, 27 Mar 2009, Jason Weber wrote:
>>
>> I haven't used it, so I don't know the details of their system.
>
> It all is explained at speedtree.com if you wish to take a look.
> I'm sure there is nothing too complex for experts.

I've paged through their site.  I didn't feel the motivation to
download their software and evaluate it in detail.

> SpeedTreeCAD is worth of looking (screenshots at speedtree.com).
> Perhaps Weber&Penn model could be extended with the curves
> you may find in SpeedTreeCAD.

My impression of the industry in general is that while fully
procedural trees are fairly easy to set up and efficient to use,
some situations often require models that are not simply realistic,
but match prescribed forms, like the particular lines of an art drawing.
They don't just want a tree of a species, they want the tree they drew.
In this case, you need modeling software like maya, developers to
add plugins, and artists to spend the time designing things.
I don't have any interest in rewriting maya or displacing blender,
so I don't think a fully custom tree design application is prudent.

Perhaps if someone could output a skeleton from a common modeling
package in an open readable format, and that might be used as the
primary level(s) from which the engine grows procedural vegetation for
secondary fill.  I've heard from people doing that.  You get the desired
silhouette without having to model every little twig.

> Open source game engine communities have hopes and wishes, but
> strong lack of experts. At this point a standalone tree generation
> and rendering software would be ok.

If you want someone to duplicate SpeedTree just to save money,
I'm not willing to expend three man years to simply duplicate
existing work.  I try to pick projects based on desirable functions
that I believe are currently missing or sufficiently lacking.

I'm a bit surprised that SpeedTree has an adequate business model
to be profitable.  The same might be said for Havok, but I think they
charge a lot more per license.  I would think any middleware
company would be wise to offer a free license to non-commercial
projects in order to hook people in.  Those presumed freeloaders could
be the same people choosing which tools to buy in their day job.

> I'm not using Ogre or any other engine. I just feel that open
> source community should have a quality game software.
> I'm happy when Linux has anything close to the editor of the
> Oblivion game. Though, I'm more interested in landscape design
> than actual gaming.
>
> If it is ok to discuss here more about the design and plans,
> I will write down my plans. I'm not sure how popular this
> list is, but this would be ok for me.
>
> Juhana

Most of FE is discussed on the dev list which is limited to dedicated
contributors.  Keep in mind that FE is a whole set of tools for many projects.
The tree dynamics engine is a relatively small part.  It just turned out to
be the only one thus far to be ready and suitable for publication.

If there are some tests and experiments you would like to try, perhaps we
could discuss where you want to go with this.  But if you have ideas that
private, proprietary, or anything that suggest signatures or lawyers, it
would be best to keep those out of a public discussion.

-- Jason Weber