Page 5 of 6

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sat Oct 22, 2016 0:49
by dpJudas
Ah, that sucks Nash. Got some nasty cold thing myself where I'm constantly coughing and freezing. Probably not the same as you're on the other side of the planet. :) No worries about the testing - this can wait.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sat Oct 22, 2016 8:16
by Nash
Did a quick test... need to shake out of this sickness somehow and stop feeling miserable...

- SSAO on models work correctly (as in, no SSAO, which I understand is expected for now)
- With gl_ssao_debug 1, depending on the player's angle, sometimes the shadow will disappear. Only happens in debug mode... in normal mode, shadows don't disappear
- gl_ssao_portals does not seem to do anything. Using Arookas' test portal WAD, I had mine set to 0 but it was still drawing SSAO across all portals. No matter what you set this Cvar to, it doesn't seem do anything functional

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sat Oct 22, 2016 20:38
by dpJudas
Thanks for testing Nash. I've committed a fix to the gl_ssao_portals having no effect (forgot to decrement it each time it did apply ssao).

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 0:51
by Rachael
Building it now. I was out today.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 1:13
by dpJudas
Thanks Eruanna. Much appreciated. :)

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 1:52
by Rachael
No problem. :)

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 6:55
by Nash
Just tried the latest version. I can't seem to find any more bugs or oddities. Everything works and looks beautiful now. =D Thanks dpJudas!

Spoilered for off-topic-ness
[spoiler]OH and by the way... this isn't really a bug report but... is there any way to make the AO not disappear near the edges of the screen? It looks good with Doom's diminished lightmode (software) but for the standalone game I'm making, it's an intentional aesthetic choice that I enforced the "dark" lightmode (Doom lighting just does not make sense for my game). So the issue here is the AO disappears when you walk too close to a wall, making it look like the player is a moving lightsource, not unlike the Doom "software" lightmode...[/spoiler]

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 13:58
by dpJudas
Nash wrote:OH and by the way... this isn't really a bug report but... is there any way to make the AO not disappear near the edges of the screen?
Technically the problem here is that around every pixel the SSAO samples nearby pixels (lets say nearest 16 pixels in X and Y). At the edge of the screen it runs out of pixels to sample from and that makes the SSAO disappear at the edges. It can be solved by making the scene framebuffer/texture larger than the final output, but due to the way the scene is being rendered right now (particular its 2D weapon) that is a little tricky to do.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 17:07
by Nash
Ah, okay. It's no big deal, I can live with it. =) Everything appears to be all good for real now! Played around with it some more and really couldn't find any more bugs.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Oct 23, 2016 23:27
by Rachael
Thank you for helping to test that, Nash! :)

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sat Nov 05, 2016 17:43
by Nash
dpJudas wrote:
Nash wrote:OH and by the way... this isn't really a bug report but... is there any way to make the AO not disappear near the edges of the screen?
Technically the problem here is that around every pixel the SSAO samples nearby pixels (lets say nearest 16 pixels in X and Y). At the edge of the screen it runs out of pixels to sample from and that makes the SSAO disappear at the edges. It can be solved by making the scene framebuffer/texture larger than the final output, but due to the way the scene is being rendered right now (particular its 2D weapon) that is a little tricky to do.
I just wanted to share that after more than 200 hours of playing Fallout 4, I only started to notice this exact effect happening in Fallout 4 last night. The effect is hilariously almost identical between F4 and GZDoom. XD

Goes to show that when you're so invested in a game, even the most hardcore nitpickers like me won't notice tiny insignificant shader weirdnesses... :P

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Nov 06, 2016 2:46
by dpJudas
It is all about attention. It is like those silly magician TV programs where they show that you can have a guy in a bunny suit walk across the scene in the back without noticing, if you just manage to keep the viewer focused on something else in the front. :)

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Nov 06, 2016 9:19
by Graf Zahl
How ready is this? There hasn't been any activity lately but so far I haven't seen anything that would prompt me to merge.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Sun Nov 06, 2016 10:33
by dpJudas
The status right now is that the branch/PR has no known issues (at least, it survived the testing Nash did) and is ready to be reviewed again.

Re: Ambient occlusion and dynlight shaders with point light math

Posted: Mon Nov 07, 2016 4:11
by dpJudas
By the way, I'm not sure if I understood your posting correctly. If you meant there are some blockers that makes you not want to merge it in its present form, then what are those?