duzenko
-
Posts
4149 -
Joined
-
Last visited
-
Days Won
50
Everything posted by duzenko
-
I'm kinda reluctant on the idea of adopting their RT technology until it's cross-vendor at least (and ideally opengl-compatible). Not mentioning the 20XX card user share.
- 23 replies
-
- 1
-
-
- nvidia
- raytracing
- (and 8 more)
-
Sorry, I am struggling to see what your point is, exactly? Also, by "newer" you mean "never"?
-
The portal in BCD closes too early as seen with r_showportals 1. Is it worth investigating? BTW wow that mission is heavy...
-
Sorry, don't understand See attached
-
You could put all in one texture and use a texture matrix for each part of it if I understand this discussion. -
Please start with the lightning fix and then we can think about the rest
-
It seems to be broken by one of these r7629 Reverted some of my previous revert in tr_render.cpp, cause it breaks foglights and shadows with the way it handles things now sigh. --------------------- r7628 Revelator: Further reverts for duzenkos code, missed one in tr_render.cpp --------------------- r7627 revert changes to duzenkos FBO code, tbh i have no idea how it got changed in the first place (gremlins ?) --------------------- @revelator, would you go through you changes and find which one caused it?
-
You mean the ambient light turns off?
-
I can confirm the blackout under water when SS is on. Probably not related to shader changes. Narrowed down to svn rev 7603. Svn rev 7631 fixes it for me.
- 1127 replies
-
- 1
-
-
- idtech4
- development
- (and 1 more)
-
The multilight shader was new files, not a change to the existing shaders. The last changes to the shaders used by the trunk code were 1) texture matrix (manifests as surfaces filled with same color if shaders out of sync with code) 2) SS micro-optimization, no visible changes expected 3) low-res stencil shadows (again, no visible changes expected)
-
This is how Crystal Grave looks to me with latest svn assets and code. What are your AA and gamma settings? BTW AA is causing a lot of error spam in console to me.
-
Svn rev 7626, partially reverts the previous commit for deforms and subviews. To test: view inside shadow volume.
-
After some thoughts, I'd rather go the BFG way but can you point me where they copy their ambientCache for ambient and interaction surfaces? Search on github did not show any. NM, found it. The BFG style is great that it does not need full copy of a srfTriangles_t for each surface and it the backend does not need to load srfTriangles_t content from a pointer.
-
It's one geo, and backendGeo is a mere pointer to that. (So that I could push the code sooner for review without a ton of changes in the backend). BackendGeo is going to disappear after code review. Effects how I see it: one geo is less confusing than two. Fewer memory writes (It's micro-optimization weekend BTW lol). One less pointer lookup in the backend if geo is struct member, not a pointer. On the other hand we might as well try to copy the ambientCache, shadowCache, indexCache, etc.
-
Try the 64 bit executable
-
Yes, please restore the FB scissor thing and make sure that r_softShadowsQuality -NN has effect (pixelated shadows). It was added back in July.
-
I think I figured out how to get rid of the two geo's in the drawsurf struct. Please see svn rev 7625 and advise. The idea is to not store frontendGeo at all since it's not used anywhere*. I tested this in TD2: chalice of kings where the guards used to visibly flicker with my earlier unsuccessful attempts. *Deforms and subviews need separate testing.
-
Can you explain why this code was deleted from FB_ToggleShadow in svn rev 7618? if(r_shadows.GetInteger() == 1) if ( on ) { if ( r_softShadowsQuality.GetInteger() < 0 && shadowOn ) shadowResolution = r_softShadowsQuality.GetInteger() / -100.; else shadowResolution = 1; qglViewport( 0, 0, glConfig.vidWidth * shadowResolution * r_fboResolution.GetFloat(), glConfig.vidHeight * shadowResolution * r_fboResolution.GetFloat() ); FB_ApplyScissor(); } else { if( primaryOn ) qglViewport( 0, 0, glConfig.vidWidth * r_fboResolution.GetFloat(), glConfig.vidHeight * r_fboResolution.GetFloat() ); else qglViewport( 0, 0, glConfig.vidWidth, glConfig.vidHeight ); FB_ApplyScissor(); }
-
Do we want to add this to the simple ambient? Or remove the simple ambient option (I'd prefer the latter)?
-
I can't comprehend the commit 7620. It's full of formatting changes like this. - basePath = Mem_CopyString(buff.c_str()); + basePath = Mem_CopyString( buff.c_str() ); Any attempt to find actual changes is going to turn to a waste of time. When I did a bulk formatting change by mistake in 2016 I got a slap on the wrist from @grayman.
-
Sorry, I still can't see it. I started the Gate house and passed a couple of doors and it just works for me.
-
How do I repeat it? I don't see this problem in e.g. Old Habits 2.
-
That needs to be replaced with qglDisableVertexAttribArray( 8 ) but I don't think it's related to invalid geometry.
-
Start with less AA and SS
-
You understand that it leads to numerous unnecessary merge conflicts in branches other than trunk?
