sandey/blender
1
0
Fork 0
forked from bartvdbraak/blender
Code Pull Requests Activity
46,836 Commits 2 Branches 0 Tags 753 MiB
fca2058924
Commit Graph

415 Commits

Author SHA1 Message Date
Gaia Clary
92b6365e21 fix: #34871 Cycles/CUDA/sm_35: Build problems probably due to issues with float3 operators (on windows?) 2013-04-06 11:52:40 +00:00
Brecht Van Lommel
3863e72c73 Fix #34679: cycles image texture alpha fringes. New rule is now that color output 
will not give straight RGB values if the alpha output is used, so that mixing with
a transparent BSDF gives the correct result.
 2013-04-05 16:43:53 +00:00
Brecht Van Lommel
54363910bc Attempt to fix #34871: random CUDA 3.5 compile failure due to recent change. 2013-04-04 23:52:33 +00:00
Brecht Van Lommel
4c8e70d049 Fix #34700: orthographic camera DOF was still not working correct. 2013-04-03 17:32:30 +00:00
Brecht Van Lommel
743552ff2a Fix #34852: multilayer SSS material rendering different in progressive and 
non-progressive integrator.
 2013-04-03 16:12:13 +00:00
Brecht Van Lommel
5c74e6dae2 Cycles: small code cleanup + fix SSS closure mixed with other closures doing 
a bit too much work.
 2013-04-02 16:37:28 +00:00
Brecht Van Lommel
be1a5a2328 Fix 34831: SSS issue with non-progressive render, due to error in closure merging code. 2013-04-02 15:53:24 +00:00
Brecht Van Lommel
952f03150f Fix #34832: SSS render freeze with object instances. 2013-04-02 14:19:05 +00:00
Brecht Van Lommel
de9dffc61e Cycles: initial subsurface multiple scattering support. It's not working as 
well as I would like, but it works, just add a subsurface scattering node and
you can use it like any other BSDF.

It is using fully raytraced sampling compatible with progressive rendering
and other more advanced rendering algorithms we might used in the future, and
it uses no extra memory so it's suitable for complex scenes.

Disadvantage is that it can be quite noisy and slow. Two limitations that will
be solved are that it does not work with bump mapping yet, and that the falloff
function used is a simple cubic function, it's not using the real BSSRDF
falloff function yet.

The node has a color input, along with a scattering radius for each RGB color
channel along with an overall scale factor for the radii.

There is also no GPU support yet, will test if I can get that working later.

Node Documentation:
http://wiki.blender.org/index.php/Doc:2.6/Manual/Render/Cycles/Nodes/Shaders#BSSRDF

Implementation notes:
http://wiki.blender.org/index.php/Dev:2.6/Source/Render/Cycles/Subsurface_Scattering
 2013-04-01 20:26:52 +00:00
Brecht Van Lommel
40b05d364e Cycles: code refactoring to add generic lookup table memory. 2013-04-01 20:26:43 +00:00
Campbell Barton
d15d78a33a style cleanup: osl and NULL pointer use, also correct sequencer gap operator id's 2013-03-27 20:27:07 +00:00
Thomas Dinges
535253cad0 Cycles / SunSky: 
* Use already calculated theta^2 for theta^3 calculation.
 2013-03-27 19:44:25 +00:00
Thomas Dinges
bf2c46f5bd Code cleanup / Cycles: 
* Remove unused variable from noise_wave().
 2013-03-27 19:08:29 +00:00
Brecht Van Lommel
5dbe5fc496 Fix #34700: cycles depth of field now works with orthographic cameras too. 2013-03-21 02:38:11 +00:00
Thomas Dinges
56538ea685 Cycles / OSL: 
* More small improvements: return immediately, and use "else if".
 2013-03-10 00:42:47 +00:00
Thomas Dinges
5ac628fb47 Cycles / OSL: 
* Avoid some conditional branches for the Noise texture and return immediately.
 2013-03-10 00:11:52 +00:00
Thomas Dinges
51f22e639e Code cleanup: 
* Cycles: Removed leftover include of "kernel_qbvh.h", which was removed in r51352.
 2013-02-21 21:05:31 +00:00
Brecht Van Lommel
909d64079a Fix #34226: cycles shadow pass got incorrectly influenced by world multiple 
importance sampleing.
 2013-02-13 16:46:18 +00:00
Brecht Van Lommel
9307565269 Attempted fix for issue with latest CUDA kernel for 32 bit windows/linux/mac, 
there seems to be some sort of compiler bug in CUDA toolkit 4.2, uninlining a
few functions seems to avoid it.
 2013-02-11 22:41:11 +00:00
Thomas Dinges
8011cbc751 CUDA / sm_20: 
* Enable Lamp MIS again, it works on a GTX 570 (which is sm_20) with latest SVN. Tested by vitos1k in IRC:.
Feedback from GTX 580 users would be appreciated on this.
 2013-02-06 23:19:14 +00:00
Brecht Van Lommel
7ecc9cfdf4 Fix non-progressive lamps with multiple samples not giving correct intensity after 
recent fix.
 2013-02-05 13:33:24 +00:00
Thomas Dinges
13499c6240 Cycles UI Code: 
* Some more variable cleanup for old texture output.
 2013-02-04 18:50:09 +00:00
Brecht Van Lommel
7c9d993347 Fix cycles intersection issue with overlapping faces on windows 32 bit and CPU 
without SSE3 support, due to 80 bit precision float register being used for one
bounding box but not the one next to it.
 2013-02-04 16:12:37 +00:00
Thomas Dinges
e3784a27d6 Cycles / OSL: 
* Fix for r53689, there are two noise types, signed and unsigned.
Caused Musgrave Texture to render differently compared to SVM backend.
 2013-02-04 11:23:40 +00:00
Brecht Van Lommel
b31d9c6cd0 Fix #34087: cycles shadow pass not properly normalized for non-progressive integrator 
with > 1 samples for a lamp.
 2013-02-03 13:10:56 +00:00
Brecht Van Lommel
736d7810a0 Fix an issue with the new cycles lamp MIS option not working correct with 
multiple lamps.
 2013-02-01 18:54:24 +00:00
Thomas Dinges
d319d68c65 OSL 1.3x / Windows 64: 
* Added missing define into SConscript file. This also makes r54232 obsolete, so removing this again.
 2013-01-31 17:38:09 +00:00
Brecht Van Lommel
122a2084b5 Attempt to fix cycles OSL link error on windows 64 bit. 2013-01-31 13:44:03 +00:00
Brecht Van Lommel
f6f5d17a50 Fix #33984: cycles shadow pass problem with CUDA. 2013-01-30 17:04:51 +00:00
Brecht Van Lommel
dc0f4b5618 Cycles: make multiple importance sampling for lamps an option per lamp now, 
disabled by default for backwards compatibility.
http://wiki.blender.org/index.php/Doc:2.6/Manual/Render/Cycles/Integrator
 2013-01-30 15:57:15 +00:00
Brecht Van Lommel
543e06ce0e Fix #33915: tweak self intersection epsilon for motion blur a bit further, still 
had some cases where there were artifacts. Also fix rendering error with shutter
time set to 0.
 2013-01-23 16:56:02 +00:00
Thomas Dinges
59b096ed00 Cycles: 
* Disable Lamp MIS on sm_20. Still enabled for CPU and sm_21 and above.
 2013-01-22 14:49:52 +00:00
Stuart Broadfoot
4c67b23c8d Cycles Hair: Fix of cardinal curve intersection routine 
Fixed some of my own mistakes in the cardinal curve intersection routine.
 2013-01-21 00:58:14 +00:00
Campbell Barton
f414d2980a code cleanup: remove some paranoid checks which would have crashed anyway earlier on. 
Also some minor formatting.
 2013-01-20 14:50:50 +00:00
Brecht Van Lommel
486abf01c6 Fix CUDA compile after last commit. 2013-01-15 20:13:19 +00:00
Stuart Broadfoot
3373b8154b Cycles Hair: Introduction of Cardinal Spline Curve Segments and minor fixes. 
The curve segment primitive has been added. This includes an intersection function and changes to the BVH.

A few small errors in the line segment intersection routine are also fixed.
 2013-01-15 19:44:41 +00:00
Brecht Van Lommel
0967b39be1 Fix #33838: light render passes for non-progressive integrator were not correct. 2013-01-15 19:17:51 +00:00
Thomas Dinges
f146317b09 Cycles: 
* CUDA: Make it more clear that sm_12 and below is not supported.
* OpenCL: __KERNEL_SHADING__ was declared twice for nvidia opencl device.
* Some reshuffle of defines in kernel_types.h. No functional changes.
 2013-01-15 19:02:17 +00:00
Brecht Van Lommel
6adfd91657 Fix #33830: cycles normal mapping was not quite correct, was not correctly 
respecting the assumption that normal and tangent are interpolated without
normalization.
 2013-01-15 16:35:05 +00:00
Brecht Van Lommel
8869c135e3 Cycles: enable lamp multiple importance sampling on the GPU, seems to be 
working ok with sm_20 and sm_21.
 2013-01-14 17:52:11 +00:00
Sergey Sharybin
36f44c6e02 Change some big functions from __device_inline to __device, which 
makes CPU kernel compilation much faster when using MSVC.

Sideeffect of this change is that CPU rendering is few percent
faster now. CUDA rendering is the same speed.
 2013-01-14 17:30:20 +00:00
Brecht Van Lommel
6cdbd1b1aa Fix #33868: cycles sample as lamp for world background not rendering with the 
correct intensity on the GPU.
 2013-01-14 16:24:47 +00:00
Brecht Van Lommel
0deb074f1c Fix #33824: cycles non-progressive render mode did not do correct path termination, 
leading to too much noise when the min bounce setting was lower than max bounce.
 2013-01-10 19:28:16 +00:00
Brecht Van Lommel
057f0a63d4 Fix #33821: cycles background intensity fix had uninitialized memory usage 
for area lights.
 2013-01-10 17:37:26 +00:00
Brecht Van Lommel
04754cc112 Fix too bright result in background multiple importance after recent changes. 2013-01-10 12:35:09 +00:00
Brecht Van Lommel
8cf374d401 Cycles: different fix for perlin noise generating nan values, now check for 
the result to be finite afterwards which is a bit faster and works for OSL
too without needing to slow down OSL itself.
 2013-01-09 22:06:03 +00:00
Brecht Van Lommel
ad10cbf04a Cycles: multiple importance sampling for lamps, which helps reduce noise for 
big lamps and sharp glossy reflections. This was already supported for mesh
lights and the background, so lamps should do it too.

This is not for free and it's a bit slower than I hoped even though there is
no extra BVH ray intersection. I'll try to optimize it more later.

* Area lights look a bit different now, they had the wrong shape before.
* Also fixes a sampling issue in the non-progressive integrator.
* Only enabled for the CPU, will test on the GPU later.
* An option to disable this will be added for situations where it does not help.

Same time comparison before/after:
http://www.pasteall.org/pic/show.php?id=43313
http://www.pasteall.org/pic/show.php?id=43314
 2013-01-09 21:09:20 +00:00
Stuart Broadfoot
ec33cacc62 Added vertex color attributes (currently limited to one) and UVs included for triangle mesh hair. 
I have also included a small speedup for the intersection test.
 2013-01-04 12:44:38 +00:00
Brecht Van Lommel
e1004eb139 Fix CUDA build error after last commit, and remove some comments. 2013-01-03 13:18:35 +00:00
Brecht Van Lommel
bf25f1ea96 Cycles Hair: refactoring to store curves with the index of the first key and the 
number of keys in the curve, rather than curve segments with the indices of two
keys. ShaderData.segment now stores the segment number in the curve.
 2013-01-03 12:09:09 +00:00