maantas

Hello everyone. finally i found a forum for 3d trees
so i am a student, and i have recently finished my internship in one company working on 3d animation film. my task was to model trees, so i spent 3 months for making 7 trees. the trees are to be used in close ups, in cinema resolution (4000px)
now i am writing report about internship, and i have some questions for you, fellow tree makers
---------------------------
well my main question is do you think 3d generators can be and should be improved, which are the areas that 3d generators do lack power in. what should be addressed and improved, what new features are needed in your eyes.
well i am coming from 3d animation (not architecture viz) field, so of course i have more "needs" for realistic trees. so if you have any ideas, wishes, even if it sounds not realistic, please post your opinion.
i am considering to write my master thesis on a topic so i need some opinions here
ps. i have just read thread "procedural tree tests" by Ernest Burden,
i d like to hear your opinion, your work looks interesting, i am planing on something similar (different methods)
pss. DennisB, in thread "My Trees" by AirWolf, you wrote, "If you ever move into L-system stuff, let me know." would there be a possibility to talk about lsystems privately or?

DennisB

Of course Maantas. I'd be happy to talk with you. I resist posting my email address in public forums anymore despite the best efforts to block spam spiders and bots. So, please go to my site and use the email link in the menu frame and send me an email that way.

Looking forward to hearing from you.

Dennis @ DigArts
http://www.gardenhose.com

maantas

ok, if anyone still reads thuse old posts,
i finished my thesis and start to post bits and peaces of to my blog.
I would realy aprishiate coments from users of vue xfog and all others.
so if anyone is interested here is a link:
*(seems i cant paste links in this forum so i will try to "encrypt" a link = you shuld remove sign"}" )
h}tt}p:/}/3dideas.wordpress.c}om/
lets hope it works
, best,
mantas

http://3dideas.wordpress.com

__________________
Mantas. some of my works here:
http://isnm.de/~talmant/works.html

DennisB

Hiya Mantas.

Nice work. I'm not a Vue user at present and most everything I'm doing with trees involves texturing techniques with standard objects (cylinders, planes, etc) and deformed meshes for realtime 3D space. I just have no experience with higher poly counts like you guys are doing. I have seen some amazing stuff though, and I'm excited by the potentials I've encountered, rudimentary forms though they are.

One of my oldest customers in France recently moved to the PS version of my main product and quickly posted a Vue tutorial on a texturing technique called Texturing for 3D Trees. It's pretty cool and translates to most any mainstream 3D application. I work in 3D where bandwidth is critical. So, it's mostly too heavy an approach for me, at least with small trees, but I've done something similar with larger trees however.

Then there's grasses, which I'm exploring now. Some of the fields tests are really quite interesting, using mixes of different wild grasses as a horizontally seamless tile. Low poly count and very nice effects as seen below (midday, sunrise and sunset). These are little 256x 512 reductions from larger screen captures, but you get the idea. The textures are 512s reduced from 2000 to 3000 pix originals. So, high rez textures are easier.

To offer an opinion about your larger question, and I don't use tree generators, I think deformed meshs have great potential, so long as you know how to approach texture creation and have the tools to make those textures possible going in. Baking the textures can add a lot too. My two cents anyway.

Dennis
DigArts
http://www.gardenhose.com

Attached Images

	midday.jpg (69.7 KB, 33 views)
	sunrise.jpg (26.7 KB, 23 views)
	sunset.jpg (40.0 KB, 29 views)

maantas

Hello DenisB,
thank you for a comment. Actually the tutorial you sent I have already seen, it is pritty good. The thing is that for what I do such level of detail is still too small One tree I had to model was used in a film where bees and wasps are the main characters, so bump maps nor displacement maps for tree bark where not enough, i had to model it manually.. It took me a month i think, for a single tree... Sad I didn't had access to zbrush by the time, that would be a solution for that very particular task. but,
Thank you very much for your thoughts, I really aprishiate it.

mr. quizzy, thank you for editing my post and adding a link, i know its a bit out of the forum policy(links are allow after person has made 10 posts in forums or something.. ), so thank you very much.

Every day I am posting a chapter from my thesis to my blog so soon my proposal will be online.
Its actually quite simple, but in my opinion results are somewhat convincing.

here i attach a picture of trunk shapes generated with a prototype which follows my proposal:
(lets see if i can post images )

__________________
Mantas. some of my works here:
http://isnm.de/~talmant/works.html

DennisB

[quote=maantas;207355]The thing is that for what I do such level of detail is still too small One tree I had to model was used in a film where bees and wasps are the main characters, so bump maps nor displacement maps for tree bark where not enough, i had to model it manually.quote]

Yes, for close up work like that only modeling will suffice.

I thought to port some bark brushes from Forestry over to ZBrush. I got distracted along the way though. They would make modeled bark textures very easy and quite sophisticated (realistic). I imagine they'd work much the same in Mudbox. Any other applications by way of similar tools?

Dennis
http://www.gardenhose.com