Commit Graph
1625 Commits
Author SHA1 Message Date
daemyung danny jang d9d0b93306 Fix a build error (#6502) 2023-01-30 10:38:36 -08:00
Powei Feng 09248cd451 Destroy FilamentInstance's mRoot in EntityManager (#6500) 2023-01-27 16:05:42 -08:00
Powei Feng 51cbd74235 Add missing STL includes (#6496) 2023-01-26 16:07:58 -08:00
Powei Feng 45246b6109 geometry: Add TangentSpaceMesh class outline (#6476)
- Defined class with documentation in comments
 - Implemented algorithm selection
 - No actual algorithms written yet

Part of google/filament#6358
2023-01-26 14:39:48 -08:00
Ben Doherty d3956adca8 matdbg: take desktop/mobile into account for SPIR-V to GLSL translation (#6488) 2023-01-26 11:30:51 -08:00
Ben Doherty 4c82f0259d Add systrace support on Apple platforms (#6480) 2023-01-23 14:11:47 -08:00
Mathias Agopian 185c83dae2 Fix a crash on Vulkan with some Adreno drivers
We were using a specialization constant to size an array inside an
UBO interface block, which caused the crash in the Vulkan driver.

However, this specialization constant was only used to workaround a
Chrome bug on WebGL. Additionally, specialized array size inside blocks
are not fully supported in spirv.

This workaround simply replaces the specialization constant by a 
constant on Vulkan.

Fix #6444
2023-01-18 10:36:43 -08:00
Romain Guy 24e1eef56e Fix glTF volumes 2023-01-05 17:38:16 -08:00
daemyung danny jang c7c184d952 Make a small optimization on drop event at filament app (#6424) 2023-01-05 08:38:28 -08:00
daemyung danny jang 6734dc6f43 Create a path from a string view (#6406) 2023-01-03 09:36:41 -08:00
Nick FisherandMathias Agopian f8684beba2 expose joint inverse bind matrices via method on FilamentInstance (#6388)
* expose joint inverse bind matrices via method on FilamentInstance

* change const pointer refs, add asserts and update release notes

Co-authored-by: Mathias Agopian <mathias@google.com>
2022-12-20 15:19:19 -08:00
Mathias Agopian ee6612ecfa Workaround for a bug in Adreno's shader compiler
At least some adreno compilers don't like returning an element of a
UBO array that is a structure in the vertex shader.
To work this around we have to copy the each of the structure fields.

Fixes #6355
2022-12-14 15:39:18 -08:00
Mathias Agopian aa5504aad3 auto instancing wasn't applied in Settings. 2022-12-14 00:52:37 -08:00
Mathias Agopian 0d72a0904b getNormalizedViewportCoord() now returns the logical viewport coords
This is a better user-facing API because materials don't need to know
about the physical viewport.
2022-12-13 16:37:49 -08:00
Mathias Agopian b5d197dfe3 Fix unused variables in release builds 2022-12-12 12:30:33 -08:00
Mathias Agopian 4215ba6476 code cleanup
fix new code quality warnings enabled by CLion
2022-12-09 11:05:26 -08:00
Koncz Levente c2fcaf7a6b Add RGTC + BPTC support (#6320)
* Add RGTC support
* Add BPTC support
2022-12-06 11:18:35 -08:00
ViktorHeisenberger 565d385132 Use Frisvad's method to compute tangent space from normals-only mesh input (#6313)
* Change tangent basis computation for normals-only to Frisvad's method

* Update RELEASE_NOTES.md
2022-11-22 09:30:37 -08:00
Mathias Agopian f619b62877 Handle shadow borders better
All shadow maps need a 1-px border for different reasons.
The directional shadow needs a "not in shadow" border because it uses
the intersection of receivers and casters in light-space.
Other shadow types need a border because of bilinear access at the 
edges. Until now, the spot/point shadow border was handled with the
sampler's CLAMP.

Instead, now, we always generate a border and in one case we fill it
with "not in shadow" (by not rendering into it), in the other cases we
render the 2px larger shadowmap so we get completely correct bilinear 
filter.

Additionally, for PCF, DPCF and PCSS we use a large filter kernel which
accesses data outside the texture, until now this was handled with CLAMP,
but that failed (2 of the edges were not handled in the same way) when 
the shadowmap was smaller than the texture.
We fix this by clamping manually the texture coordinates.


The real motivation for these changes is to allow the use of a
shadow Atlas later.
2022-11-21 15:38:25 -08:00
Mathias Agopian 77fcf4d260 don't store the light direction in the shadowmap structure
We already have it in the Light structure.
2022-11-17 11:07:47 -08:00
Mathias Agopian 8b96af8e3f AtlasAllocator improvements 2022-11-16 09:20:22 -08:00
Mathias Agopian 641a0bda9e spell specialization correctly 2022-11-16 09:20:00 -08:00
Mathias Agopian b1e438e141 move vsm's MSAA sample count to the view settings
VSM's MSAA is no longer a "per light" setting, it is now global for
the whole view.
2022-11-14 15:29:08 -08:00
Mathias Agopian 6ec5fd86fa rework how we set the viewport for shadows
The viewport is now entirely handled by ShadowMap, that is, we're no
longer setting a viewport in the renderpass, instead the ShadowMap
uses a post projection transform to achieve the same effect.
This will become useful when store shadow maps in an Atlas.
2022-11-10 15:15:43 -08:00
Mathias Agopian c05e2ec47d improve variant filter
- variant filter didn't filter the VSM variant of depth variants

Fix #6274
2022-11-08 16:48:54 -08:00
Mathias Agopian 9768f49714 unify point and spotlight shadows
point light shadows are not sampled just like spotlight shadows, the
only difference is that we're calculating the face first to access the
corresponding shadowmap data (including the light transform).
2022-11-08 16:35:54 -08:00
Mathias Agopian 6c54cfe88a Shadowmaping code improvements
- the layer is now stored in the shadow ubo
- we now have one shadow ubo per face
- ShadowMap holds UBO index even for cascades
- move some shadow "view" uniforms into the shadow UBO
2022-11-08 15:25:33 -08:00
Mathias Agopian 2dc8d6bcc0 minor cleanups, comment fixes
- update comments
- remove unused or unneeded methods
2022-11-08 15:25:33 -08:00
daemyung jang 0a7fe52124 Trim a whitespace (#6272) 2022-11-04 08:36:36 -07:00
Ben Doherty 67babceb73 Fix backend_test to work with Metal arg buffers (#6135) 2022-11-03 11:04:29 -07:00
daemyung jang e64b498d6e Update a comment about the range of morph target weight (#6265) 2022-11-03 09:19:01 -07:00
Benjamin Doherty e9728dcc8d Fix iOS build 2022-11-01 15:03:44 -04:00
daemyung jangandRomain Guy 3b40e52c98 Recompute bounding boxes with morph targets (#6246)
* Recompute bounding boxes with morph targets

* Reduce the number of heap allocation while recomputing bounding boxes

Co-authored-by: Romain Guy <romainguy@curious-creature.com>
2022-11-01 11:20:28 -07:00
daemyung jang de8c7969f3 Return the correct permit type corresponding to the component type (#6255) 2022-11-01 11:03:27 -07:00
Ben Doherty c5cc7fe26a Metal: generate argument buffer structures ourselves (#6242) 2022-11-01 12:44:42 -04:00
Ben Doherty d04b0a5c73 Update Vulkan headers to v1.3.232 (#6256) 2022-11-01 12:38:23 -04:00
daemyung jangandMathias Agopian 1a6338569a Calculate primitive's AABB correctly (#6232)
AABB of morph target is displacement data so it should be added to
primitive's base AABB.

Co-authored-by: Mathias Agopian <mathias@google.com>
2022-10-28 11:51:02 -07:00
daemyung jang 676694e458 Fix glTF breaking issue (#6239) 2022-10-28 09:56:01 -07:00
Mathias Agopian 2ca218928f utils: get rid of GrowingSlice
it's just not used anywhere. change the implementation of Slice<> to
use two pointers.
2022-10-27 22:48:45 -07:00
Mathias Agopian 90f32e7277 A basic POT 2D allocator
This 2D allocator only supports power-of-two, square allocations. It
uses a quadtree to store the allocated regions.

The quadtree implementation currently has a fixed depth (templated).

The allocator does a best-fit match, but currently doesn't implement a
free method.
2022-10-26 11:02:17 -07:00
daemyung jang b85d9afdb3 Remove an useless whitespace (#6226) 2022-10-26 00:10:33 -07:00
Ben Doherty e75c3020c8 Initial support for Metal compute shaders / SSBOs (#6189) 2022-10-18 16:21:57 -04:00
Philip Rideout e157023fc1 Use mField style for fields in FFilamentInstance 2022-10-17 15:49:19 -07:00
daemyung jang 784416f658 Support glTF EXT_meshopt_compression 2022-10-17 09:57:58 -07:00
Mathias Agopian acd5823953 Add support for point light shadows
[experimental]
2022-10-10 13:03:36 -07:00
Mathias Agopian 734d8ff85c rename upcast() to downcast()
it was down casting all along!
2022-10-07 17:50:56 -07:00
Mathias Agopian a02ef1a6c7 Remove the limit to 14 spot shadow maps, now 64.
Spotlight shadow maps are now limited only by h/w limits (UBO size and
2d array maximum layers), which works out to about 146 shadows.

In practice we hardcode the limite to 64, but this can be changed at
compile time. A higher maximum increases memory usage.

There used to be a bitfield per renderable with a bit for each shadow
map indicating if the renderable was "visible" from this light. The
limit to 14 came from that bitfield. All shadow maps used to be culled
first, setting the bitfield appropriately.

Instead, we are now processing spotlight shadows one at a time,
performing culling as we go. For this reason we don't need a bit per
shadow map.
2022-10-07 14:07:51 -07:00
Mathias Agopian 44fdf83e4c fix multiview mode in gltf_viewer
The "shadow" camera now works again.
2022-09-30 16:12:15 -07:00
Philip Rideout 98e4cdbd4c gltfio ubershader mode: fix re-loading.
Fixes #6128
2022-09-30 13:54:36 -07:00
Mathias Agopian 5aa414dd24 support for ELVSM and 32 bits xVSM
This basically adds two settings:
- highPrecision at the View level, which controls the bit depth
  of the vsm shadow texture used. This affects all shadowmaps.
- elvsm per shadowmap, which enables Exponential Layered VSM.

the main change in the shaders is that we now always output the 
"negative" EVSM, even when it's not enabled. If no shadowmap uses
ELVSM, then the texture is stil an RG texture and the calculation is
lost (with some luck culled by the shader compiler), either way it's
not a lot of math and it's done only once per shadow texel.

During the color pass, on the other hand, we compute the "negative"
EVSM only if enabled.
2022-09-30 12:04:27 -07:00