the only time that errors do occur, from what I''ve found so far, is when two triangles join and the edge shares the same ordering. But this is almost always caused by a bad model, and I can put in a fix. (simply treat it as two seperate edges)

quote:

---

no... because even a single triangle floating in space will still have a looping silhouette edge..

---

Do you mean that each model have a unique looping silhouette edge ?
It''s true for convex models, but not for concave models.

heh no..
there is almost always more than 1.

Hey guy!!!

That''s all right here! Works perfectly

-----------

Pentium Celeron 466MHz
Running in Windows 2000 Pro
ATI Rage 128 AGP - XPert 99/00
192Mb RAM

-----------

And, how you can read the MS3d file compressed?? It''s very good the loader and lights

Rafael G. Donato

ATI Rage fury pro (rage 128 pro chip) 32 MB
runs perfectly!

I hope you will release a demo later...

quote:

---

there is almost always more than 1.

---

Then how can you optimize the algorithm with looping silhouette edges ?
Since you don''t know the number of loops, the only acceptable algorithm is to test all edges, isn''t it ?

runs well on my radeon 8500.

i have a question. why aren't all the tri's casting shadows? is it because you do it by-elements, such as head, body, etc. because i never see the helmet cast any shadows on the guy's face.

edit: actually, it does cast shadow on his face, i just didn't notice. you're right, the whole don't-darken-but-light-by-some-ambient-light thing works really well, and looks awsome (assuming that blue light is the ambient light). but sometimes it could be a little bit confuzing.

and another question. you used the stencil buffer, right? what color, depth and stencil bits did you use?

---
shurcool
wwdev

[edited by - shurcool on July 24, 2002 9:29:19 AM]

Runs very slow on my Ati RadeOn 8500 - Windwos XP again I think.
Drivers are okay (I''m using omega drivers)

Runs just fine, nice work !

P4 1500
256mb ram
ATI Radeon 7200 64mb
WinXP

[edited by - Crawl on July 24, 2002 1:40:24 PM]

quote:

---

Original post by shurcool
runs well on my radeon 8500.

i have a question. why aren''t all the tri''s casting shadows? is it because you do it by-elements, such as head, body, etc. because i never see the helmet cast any shadows on the guy''s face.

and another question. you used the stencil buffer, right? what color, depth and stencil bits did you use?

---

the blue ambient light doesn''t cast shadows as it is right now.. (which does look ''bad'' compared )
I''d imagine it''d be pretty difficult to not do stencil shadows without the stencil buffer (unless you did some crazy alpha buffer crazyness ) but my hardware is so old I wouldn''t dare attempt that. depth = 24bit, stencil = 8 bit. 1 bit stencil makes complex concave objects pretty much impossible to shadow...

quote:

---

Original post by vincoof
Then how can you optimize the algorithm with looping silhouette edges ?
Since you don''t know the number of loops, the only acceptable algorithm is to test all edges, isn''t it ?

---

I never said it was a faster algorithm as far as cpu usage goes. it''s no faster that way. But as far as what can be optimized later on when actually drawing the thing, then it is significantly faster.