diff options
| author | Clément Foucault <foucault.clem@gmail.com> | 2018-11-19 20:03:06 +0300 |
|---|---|---|
| committer | Clément Foucault <foucault.clem@gmail.com> | 2018-11-19 20:05:15 +0300 |
| commit | 7bb512594cd9502fea290ac6124f2eb5fd3cfce8 (patch) | |
| tree | 190b9e4184e6eabf3349e74bf6d58c6da9f29288 /source/blender/draw/engines/workbench/workbench_materials.c | |
| parent | 2b56d2183972a0cb3b1355652a6709599ae6af0d (diff) | |
Workbench: Use non-negative lighting evaluation
This makes the lighting a bit more diffuse but don't produce negative
values.
Add a bias of 1.5f to make the lighting a bit more directionnal.
The implementation is based on:
https://github.com/kayru/Probulator/blob/master/Source/Probulator/SphericalHarmonics.h#L136
which is derived from:
http://www.geomerics.com/wp-content/uploads/2015/08/CEDEC_Geomerics_ReconstructingDiffuseLighting1.pdf
The shader implementation is optimized and has the same runtime cost
as previous method:
* no sh eval : 0.13ms
* prev sh eval : 0.14ms
* new sh eval : 0.22ms
* new sh eval opti : 0.14ms
Diffstat (limited to 'source/blender/draw/engines/workbench/workbench_materials.c')
| -rw-r--r-- | source/blender/draw/engines/workbench/workbench_materials.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/source/blender/draw/engines/workbench/workbench_materials.c b/source/blender/draw/engines/workbench/workbench_materials.c index 14b4c5e6f4c..572a26c86bd 100644 --- a/source/blender/draw/engines/workbench/workbench_materials.c +++ b/source/blender/draw/engines/workbench/workbench_materials.c @@ -92,7 +92,7 @@ char *workbench_material_build_defines(WORKBENCH_PrivateData *wpd, bool use_text } BLI_dynstr_appendf(ds, "#define STUDIOLIGHT_SH_BANDS %d\n", STUDIOLIGHT_SH_BANDS); - BLI_dynstr_appendf(ds, "#define STUDIOLIGHT_SH_MAX_COMPONENTS %d\n", STUDIOLIGHT_SH_EFFECTIVE_COEFS_LEN); + BLI_dynstr_appendf(ds, "#define STUDIOLIGHT_SH_MAX_COMPONENTS %d\n", WORKBENCH_SH_DATA_LEN); str = BLI_dynstr_get_cstring(ds); BLI_dynstr_free(ds); |