| Age | Commit message (Collapse) | Author |
|---|
|
Limit influence of refraction scale to avoid pulling in lots of background 'deep blue' pixels never written in the distortion map pass.
|
|
|
|
This is more or less to achieve the non-ALM bloom behavior in most circumstances.
Why scale it? Normalized blinn-phong gets brighter the higher the specular exponent. As it stood, the old spec contribution was extremely bright despite looking "correct". This is largely eyeballed, but generally looks better for all light sources applied to the surface of the water.
|
|
The atomospherics were being applied both in the G-buffer and during the sun light pass. This ensures that is not the case.
Additionally, re-enable specular on the water plane, scale the bloom factor to something more reasonable, and also someone broke specular on emissive surfaces. That's fixed too now.
|
|
Fix reflection/distortion map culling planes again.
Fix broken handling of shadow disables in ALM forward shaders.
|
|
Fix up culling issues from perf work and fix Depth of Field rendering to get depth values properly.
Baseline for performance work.
|
|
Fix distortion map rendering in deferred mode not including underwater fog effects.
Fix distortion map rendering not including post-deferred content at all.
Fix distortion map rendering not including anything but sky when camera is underwater.
Update sun_up_factor/sunmoon_glow_factor uniforms even when sun disc isn't in use.
|
|
Fix bug with setting texture matrix for rigged mesh.
Fix z-fighting between sea and sky in ALM by cheating sky, stars, and moon using gl_FragDepth.
Fix handling of atmospheric haze glow w.r.t independent sun/moon positioning
(we can no longer depend on them being mutex to each other).
|
|
|
|
Timmy).
|
|
Use a new edge water shader to allow forcing frag depth to avoid z-fighting at back edge of water edge pieces.
This will not work for ALM because forcing the depth breaks the use of the depth to backproject gbuffer position which breaks lighting calcs.
|
|
Prepare for using ifdef to conditionally engage using gl_FragDepth in water shaders
(because GL requires that writing to gl_FragDepth on any path means you need to
write a depth value on every path).
|
|
Force frag depth in water shaders to avoid z-fighting at back edge of water edge pieces.
|
|
Modify distortion vec calc in shader to account for modified range of reflection scale above/below.
|
|
|
|
next textures are identical.
|
|
|
|
|
|
|
|
Eliminate 20+ callsites with copy-paste of getPosition and/or getNorm code.
Make pipeline use getShadowTarget/releaseShadowTarget consistently.
|
|
Make shaders use consistent naming and parameter order for transport and atmospheric helpers.
Share transport and gamma correction code where possible.
Add lots of asserts and other validation for when things don't go as planned.
Engage dumpShaderSource to get more source output with shader compilation fail.
|
|
Prepend dir names to comments in water shaders to differentiate them in nSight captures.
Fix calc of reflection camera modelview mats to properly reflect the environment.
|
|
Make stars use higher alpha and brighten reflected color before mixing to get stars to show up in reflections on water.
|
|
for current to next blending.
Modify LLSettingsFoo::buildDefaultFoo to use a static and avoid re-validation of default sky/water/daycycle settings.
Remove all references to gSun/MoonTextureId globals (they should come from sky settings now).
|
|
format.
|
|
and unify on using the version that works on OSX.
Add more logging to shader loading about fallbacks and loading succcess.
Add frag shaders for sharing sRGB and normal encode/decode via GL shader linkage.
|
|
viewer
FIXED
|
|
|
|
|
|
|
|
|
|
10.6.8 nV driver issues
|
|
conversion in one place
|
|
|
|
issues, alpha-blend objects still missing haze underwater
|
|
|
|
reflect/distort RTs
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Reviewed by Graham
|
|
|
|
|
|
errors, and shiny environment map errors due to lack of negative Z)
|
|
|
|
for texture indices in the data stream. Also rework gl_FragColor overrides to not collide with some odd driver implementations.
|
|
|
