diff options
| author | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2014-06-20 23:21:05 +0400 |
|---|---|---|
| committer | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2014-06-22 00:31:44 +0400 |
| commit | 8fbd71e5f2862e6efd0e3bcb015f1f6201ccca14 (patch) | |
| tree | 0d889efc32554d0eb4ad6f8b1f46b4885b19991d /intern/cycles/render/shader.h | |
| parent | 88d8358f91e1306a67d59250162443194ee9edcf (diff) | |
Cycles: improved Beckmann sampling using precomputed data
It turns out that the new Beckmann sampling function doesn't work well with
Quasi Monte Carlo sampling, mainly near normal incidence where it can be worse
than the previous sampler. In the new sampler the random number pattern gets
split in two, warped and overlapped, which hurts the stratification, see the
visualization in the differential revision.
Now we use a precomputed table, which is much better behaved. GGX does not seem
to benefit from using a precomputed table.
Disadvantage is that this table adds 1MB of memory usage and 0.03s startup time
to every render (on my quad core CPU).
Differential Revision: https://developer.blender.org/D614
Diffstat (limited to 'intern/cycles/render/shader.h')
| -rw-r--r-- | intern/cycles/render/shader.h | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/intern/cycles/render/shader.h b/intern/cycles/render/shader.h index 84be4b469d8..0ed6d2ddf01 100644 --- a/intern/cycles/render/shader.h +++ b/intern/cycles/render/shader.h @@ -149,6 +149,7 @@ protected: AttributeIDMap unique_attribute_id; size_t blackbody_table_offset; + size_t beckmann_table_offset; }; CCL_NAMESPACE_END |