Speaking for Gilles (I must be mad these clothes are too big for me, they
don't fit, and I can't find a good scaling macro...)

Well, I really don't believe Gilles doesn't know how to scale a tree, but he
says that you can have the need for two trees of different heigh but with
the same leaves, if they are both scaled down to the unit box, you'll spend
you life trying to find the good scaling factor to get them having the same
leaves (same problem with the texture... ) the different trees *needs* to be
consistent, the don't need to be the same size)

Philippe



>
>Sorry to disagree again, but no, It can be done with a good uniform scaling
of
>the object so it keeps the curent scale of the objects. This should be
quite
>easy to implement.
>something like this is in my mind..
>
>translate -1*MinExtent
>//that should center the object with one corner on <0,0,0>
>
>Then there should be a scaling, something like:
>scale 1/(MaxExtent-MinExtent)
>...
>in my mind the tree will keep it's scale in sizes, but be downscaled to
fit...
>perhaps I'm to tired to think straight, but then, what good is straight?
>(Go chaos, go)
>
>
>
>Gilles Tran wrote:
>>
>> Hmm. It could be useful in some cases, and a little cumbersome in others.
>> Personnally I prefer having the trees not scaled and adjust the size
using the
>> dummy bounding box: for instance, if you generate several trees just but
changing
>> the random seed or the recursion level, you'll get trees of different
sizes but
>> with a similar scale. If the trees were post-scaled, you'd get even-sized
trees but
>> with actually different scales, so you would have to make trial-and-error
scaling
>> to have trees consistent with each other. There's also the fact that the
trees
>> don't exactly fit in a box. Some of them crawl, some shoot up, so there's
no ideal
>> unit-size bounding box. In fact people who provide really good 3D trees
also
>> provide the "real life" bounding box : see
http://iris8.cirad.fr/index.htm and look
>> at the plant catalogue.
>> Now if you really want it, I can add it as an option.
>> Gilles
>>
>> Spider wrote:
>>
>> > hmm, this probably doesn't fit in this group, but should be in a
separate mail
>> > to Giles. but then, hehe.
>> >
>> > To ease it all up with object placing and scalability, can you make the
trees
>> > all downscaled(upscaled) to a bounding_box{<0,0,0>,<10,10,10>}. This
would help
>> > a lot, and would also remove the need to recalculate the tree to fit.
>> >
>> > I have an automated rescaler #macro here somewhere, if you are
interested...
>> >
>> > Lewis wrote:
>> > >
>> > > I had another weird idea:
>> > > Let's compile a library of trees, made with Giles macro. Using a
macro
>> > > to generate random trees, each one starts it with a different seed
and
>> > > sends in the results (the include files). It could be very useful.
Maybe
>> > > someone will need a couple of trees sometime. Instead of parsing,
say,
>> > > 100ds of them, he will just download them.
>> > > Anyone with me? If there will be enough enthusiasm I'll set up a web
>> > > site or somethin'.
>> >
>> > --
>> > //Spider
>> >         [ spi### [at] bahnhofse ]-[ http://www.bahnhof.se/~spider/ ]
>> > What I can do and what I could do, I just don't know anymore
>> >                 "Marian"
>> >         By: "Sisters Of Mercy"
>
>--
>//Spider
>        [ spi### [at] bahnhofse ]-[ http://www.bahnhof.se/~spider/ ]
>What I can do and what I could do, I just don't know anymore
>                "Marian"
>        By: "Sisters Of Mercy"

Post a reply to this message