|
|
Ok let's try to answer all the questions...
Spellcaster,
I would say this is the main interest with IOTDs: they give you that weird coding vibe, drive you nuts, and push you to code. Don't panic then, I've been jealous of countless IOTDs so far :)
tcs,
I'm afraid my coding life is way bigger :))
(www.codercorner.com/Pierre.htm)
This is just an excerpt of what was produced since the last IOTD, *grin*
The "engine" is not opensource so far. Basically this is just a generic handy framework to quickly test various algorithms [I believe it does the job! For example implementing the glow took a few hours, reusing parts of the engine already in place. That's where you love the OO way (Altair? Nick? :)]
Revolver,
There are bigger screenshots in you like it:
http://www.codercorner.com/Glow2.jpg
http://www.codercorner.com/Glow3.jpg
http://www.codercorner.com/Glow4.jpg
Sagacity,
*That* is a private joke no-one will get :)
Pants,
:) Maybe immature, but admittedly you'd better find a sexy project when it comes to mixing together so many complex things. Else you'll just give up on the first serious impossible-to-fix bug.... :)))
Denninger,
Ahahah! Speaking of immature.... :)))
I don't mind people telling me this or that sucks, but if you don't say as well *why* it sucks, you're just throwing away one more pitifully useless post among gazillions of them, aren't you?
Prophylon,
As I said: regarding milady, work in progress. There will be a demo (I hope so!) as soon as it's finished (or good-looking enough). For the moment I've just mixed the mesh-related techniques (skinning, PM, subdivision surfaces), and I still have to combine the dynamics part on top of that (mainly hair/cloth). The cloth code is ready (there's an oooooold demo here: http://www.codercorner.com/Cloth.htm) but not applied to the dress so far (no technical problem involved, I just didn't have the time). The hair part is the only real remaining challenge. If you attack the problem in the same way as cloth, you get decent results, but waaaaay to slow. So I'll have to hack my way out of it to make the whole thing realtime.
I have some old snapshots for hair. The pics are ugly (ugly model, random hair texture) but I was testing the *dynamics* here!
http://www.codercorner.com/Hair03.jpg
http://www.codercorner.com/Hair07.jpg
http://www.codercorner.com/Hair08.jpg
http://www.codercorner.com/Hair09.jpg
...yeah, I've been traumatized by the long hair in Cat's Eye when I was young... :)))
Vorteks,
As Revolver said, that's not a problem for a well-driven GeForce. Anyway 200.000 is far too high, the model does not need it since it's meant to be quite smooth. That's the problem of subdivision surfaces. You get maaaany faces even in places you don't need them. *Displaced* surfaces are way better since they give you the real thing: details! (as detail textures do). Now, for a lady's smooth skin, you certainly don't want/need bumps and displacement maps, so the number of faces should be very very smaller in the end (and reused in the hair maybe).
Revolver & Scrambled Monkey (*),
The opacity map was missing for eyelashes, I know.....
(*) Ahah, sounds like the name of your next-hype britpop band!! :)))
Max,
Right! Especially when you recompute all the normals on-the-fly...
Zed Zeek,
There's actually an opacity map missing :)
DirtyPunk,
With such a nickname, don't tell me it would stop you!
P.
|
