4 Commits over 28 Days - 0.01cph!
Cube texture creates array slice image views allowing each face to be treated as a framebuffer
Visualize depth uses screen space coords
[stash] Foundations for VK_KHR_multiview and cubemap render targets
Scene system listeners I'll need later
New shadow mapper: spot and point lights so far
* Move all logic to c# for full ownership, delete old native code
* Bindless shadow map textures, delete old shadow atlas
* Spot Lights
* Point Lights: cube textures with hw pcf poisson sampling
* Add user setting r.shadows.quality (0-3) low->high for better filtering
* Scale shadow map resolution based off of screen size the light occupies
* Debug system
Scene system listeners I'll need later
Clean up old dead code
Collect and pass around SceneLightObjects instead of the LightDesc ptr
Managed shadow mapper, entry point from lightbinner, creates and renders shadow frustums
Set gpu light data from c#, pass shadowmap bindless texture id, use it from lighting shaders
Visualize depth uses screen space coords
vfx_vulkan: SV_ViewID is a built in semantic
fucking with VK_KHR_multiview and cubemap render targets
fucking with VK_KHR_multiview and cubemap render targets
Don't return a new list each light gather, reuse the already allocated memory
Collect and pass around SceneLightObjects instead of the LightDesc ptr
Stubbed managed shadowmapper appsystem (this is such bullshit for a simple interface for scenesystem)
Managed shadow mapper stub, entry point from lightbinner
Add CommandList.Set( StringToken, Texture )
Dispose of AssetPreviews after we've rendered their thumbnail. PreviewScene.Dispose destroys it's Scene
ISceneView::GetWorldToRender & related return ISceneWorld not CSceneWorld
Clean up old dead code
Create and render shadow frustums from managed
Scene system listeners I'll need later
Remove native shadow rendering
Set gpu light data from c#, pass shadowmap bindless texture id, use it from lighting shaders
Depth bias shadows
Visualize depth uses screen space coords
