#jira na
#rb trivial
#ROBOMERGE-AUTHOR: sara.schvartzman
#ROBOMERGE-SOURCE: CL 21096155 in //UE5/Release-Engine-Staging/...
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v972-20964824)
[CL 21099282 by sara schvartzman in ue5-main branch]
* Card sampling with squared weights instead of cutoff in order to allow surface to leak into tiny areas without coverage
* Bump min static mesh LOD to reduce mismatches between surface cache and geometry
Card placement:
* Move near plane back outside of mesh bounds during capture in order to handle displacement mismatches
* Discard surfels seeing triangle back faces only when sky isn't visible
* New additive card placement, which can be used for continuous LOD. Instead of clustering, it tries to find the best split plane until the entire surface isn't covered.
* Lower voxel and min card size to be able to cover smaller areas and place more cards
* Spawn multiple surfels per cell if they cannot be merged
* Fixed surfel generation and card size
#ROBOMERGE-OWNER: krzysztof.narkowicz
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 21089854 via CL 21089869 via CL 21089871
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v972-20964824)
[CL 21096306 by krzysztof narkowicz in ue5-main branch]
+ fix internal test sub category names in Test Automation window
#preflight 62d03045dc4397d38403efb5
#jira none
#rnx
#rb Chris.Constantinescu
[CL 21091332 by Jerome Delattre in ue5-main branch]
#preflight skip
#ROBOMERGE-AUTHOR: rex.hill
#ROBOMERGE-SOURCE: CL 20986800 via CL 21080942 via CL 21081164 via CL 21081436
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v972-20964824)
[CL 21083114 by rex hill in ue5-main branch]
* Card sampling with squared weights instead of cutoff in order to allow surface to leak into tiny areas without coverage
* Bump min static mesh LOD to reduce mismatches between surface cache and geometry
Card placement:
* Move near plane back outside of mesh bounds during capture in order to handle displacement mismatches
* New additive card placement, which can be used for continuous LOD. Instead of clustering, it tries to find the best split plane until the entire surface isn't covered.
* Lower voxel and min card size to be able to cover smaller areas and place more cards
* Spawn multiple surfels per cell if they cannot be merged
* Fixed surfel generation and card size
#ROBOMERGE-AUTHOR: krzysztof.narkowicz
#ROBOMERGE-SOURCE: CL 21078868 via CL 21078893 via CL 21079371
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v972-20964824)
[CL 21082291 by krzysztof narkowicz in ue5-main branch]
#rb brian.karis
#tests a/b tested on multiple platforms to make sure no perf or memory regression; sanity checked several static meshes to make sure auto calculated screensizes stayed similar
#ROBOMERGE-AUTHOR: jian.ru
#ROBOMERGE-SOURCE: CL 21045756 via CL 21045905 via CL 21045943
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v972-20964824)
[CL 21075377 by jian ru in ue5-main branch]
* made some trace messages optional to reduce traffic and trace file size, as ParallelFor doesn't need them
* optimized out multiple task lookups per message
#preflight 62ceaa3bf123dd18360cc5c5
[CL 21069519 by Andriy Tylychko in ue5-main branch]
Initial submit, compiled out by default.
#rb Peter.Engstrom,Mattias.Hornlund,Brian.Bekich,Ryan.Gerleve,John.Barrett
#jira UE-158381
#preflight /62cd6fc7ad490b52da0515a3
[CL 21068370 by Peter Engstrom in ue5-main branch]
