| Age | Commit message (Collapse) | Author |
|---|
|
|
|
geenz/mirrors-optimization-pass-1
|
|
screen
|
|
So far 4 seems like a good balance for performance and quality. 2 is great for quality, with 6 or 8 being for higher performance. Also bring back the gaussian filter - may end up adding the FXAA filter though instead.
|
|
|
|
SimulatorFeatures
|
|
|
|
|
|
|
|
flag. Final implementation would likely be very different.
|
|
|
|
refresh has surprising performance cost
|
|
debug setting.
|
|
DRTVWR-583
|
|
|
|
|
|
|
|
above water.
|
|
|
|
drawing water twice, but needs a better long term solution.
|
|
|
|
like to read from a depth buffer that is bound for writing
|
|
|
|
|
|
|
|
Incidental fix for crash when mWaterPool is null.
|
|
LLEventPoll hack.
|
|
|
|
|
|
|
|
own passes and unify sky and water haze in forward rendering shaders.
|
|
|
|
lights rendered scale with the "Quality & speed" slider
|
|
|
|
probes more wishy-washy (removes dark splotches where probes get stuck in walls)
|
|
|
|
|
|
DRTVWR-583
|
|
DRTVWR-583
|
|
DRTVWR-583
|
|
DRTVWR-583
|
|
DRTVWR-583
|
|
DRTVWR-583
|
|
No repro, but likely cause is in mOctreeNode being NULL.
Marked octree group as dead when destroyed and added isDead verification
|
|
|
|
|
|
|
|
enabled by default
|
|
post-process buffers. Disabled by default.
|
|
DRTVWR-583
|
