Jump to content

  • Log In with Google      Sign In   
  • Create Account

Banner advertising on our site currently available from just $5!


1. Learn about the promo. 2. Sign up for GDNet+. 3. Set up your advert!


multipass polygon filling


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
No replies to this topic

#1 rAw   Members   -  Reputation: 122

Like
Likes
Like

Posted 23 November 2001 - 12:26 PM

y0w ppl! when i was coding my new,awsome multipass triangle filling pipeline (with usage of multitexturing ofcourse) :D ,i''ve noticed some stupid artifacts. the main algo is something like this: blend_mode=decal; while (is_there_any_texture_in_the_pipeline) { set_blend_mode; (depends on previous pass, but usaly it''s "add" ) render_pass; pop n textures from pipeline (where n is the # of texture units in gfx card, n<=#of_textures_in_pieline) } 1st texture in the pipeline lies at the top of stack. While the opengl api defines the way how output pixel is generated is quite simple to "emulate" that algorith using screen buffer as a temporary value. It works quite good, it''s fast, but the entire algo crashes when this "temporary value" is not 0 at the beginning (0 + 1st_texture=1st_texture ,that''s why blend mode is initialized with decal). what do i mean by that? when texture_surface (aka q3a: texture_shader) will be, lets say, added to the background , the overall effect of this loop will be a little inmathematic is there any way to omit that aritifact? (a fast one would be great!! i don''t want to create another screen buffer , play with stencil , etc)

Sponsor:



Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS