September 9, 2013, 23:46 (GMT) |
Volume: * Rename OSL closure and file to reflect recent changes. |
September 9, 2013, 23:01 (GMT) |
Volume: Cleanup and refactor of closure code in volume.h. * Remove unused code (the brute force henyey-greenstein eval and sample function and double peaked functions). Maybe we need something here in the future, but the current code just became too confusing. * Remove the CLOSURE_BSDF_DOUBLE_PEAKED_HENYEY_GREENSTEIN_ID, instead use CLOSURE_VOLUME_ISOTROPIC_ID now. The final naming of the node/closure is still unsure, but better to use one consistent naming scheme. When the "g" parameter is 0, the closure is Isotropic anyway, so maybe it's ok, we'll see. * Now we use 2 counterparts to bsdf_eval() and bsdf_sample(): volume_eval_phase() and volume_sample() ToDo: * Reflect these changes for OSL. |
September 9, 2013, 20:13 (GMT) |
OSL implementation of the Henyey-Greenstein volume closure. A lot of cleanup included. |
September 9, 2013, 14:17 (GMT) |
Volume: * Fix some render errors (no caustics etc.), the patch accidentally removed some lines in kernel_path.h. * Remove unused bsdf_eval_add() function. |
September 9, 2013, 01:34 (GMT) |
* Put some more volume code behind __VOLUME__. * Compiling on the GPU works and renders fine (sm_21 here), if you want to test uncomment the line in kernel_types.h. |
September 9, 2013, 01:02 (GMT) |
Cycles / Volume Rendering: * Initial and experimental implementation of volumetric rendering. At this stage it is already usable, although there are a lot of limitations still: - No Volume rendering in Branched Path integrator. - No Smoke volume rendering - No OSL support - UI and data structures are WIP, be careful when saving files with this. - ... Credits: * Based on the original patch by "storm". * Cleanup and improvements by Stuart Broadfoot. * A crash fix by Lukas Tönne. Thanks a lot guys! * Also some cleanup and refactor by myself. Some WIP docs and infos: http://wiki.blender.org/index.php/User:DingTo/Volume |
September 9, 2013, 00:22 (GMT) |
Branch maintenance: * Revert r59183 and r57612 of my branch. Those changes were experimental. In case we need it, it's still in SVN history. * Also fix a mismatch in gpu_shader_material.glsl. Trunk and my branch are now equal, except splash.png. Compared with Meld, thanks to Campbell for the suggestion. :) |
September 8, 2013, 23:57 (GMT) |
Merged revision(s) 59906-59929 from trunk/blender into soc-2013-dingto. |
September 7, 2013, 00:52 (GMT) |
Merged revision(s) 59751-59905 from trunk/blender into soc-2013-dingto. * Resolved a conflict in intern/cycles/render/object.cpp. |
September 2, 2013, 19:35 (GMT) |
Merged revision(s) 59605-59750 from trunk/blender into soc-2013-dingto. |
August 28, 2013, 14:48 (GMT) |
Merged revision(s) 59491-59602, 59604 from trunk/blender into soc-2013-dingto. |
August 28, 2013, 13:28 (GMT) |
Cycles / Sky Model: * Rename Albedo to Ground Albedo, and improve tooltip. * Soft UI Range for Turbidity, between 1 and 10, higher values can produce weird results. I checked and other engines clamp this to 10 as well. * Some code cleanup. |
August 25, 2013, 12:23 (GMT) |
Merged revision(s) 59438-59490 from trunk/blender into soc-2013-dingto. |
August 23, 2013, 19:13 (GMT) |
August 18, 2013, 23:54 (GMT) |
Cycles / Sky Model: * Remove the KernelSunSky struct, which limited us to 1 Sky texture for SVM. Now we can use more than 1 Sky shader, which allows some artificial effects like 2 Sun/Sky's or so. * New and old model are now working in OSL as well. Patch by Lukas Tönne, with some small tweaks by me. Thanks! ToDo: * Code cleanup! |
August 18, 2013, 20:17 (GMT) |
Merged revision(s) 59185-59246 from trunk/blender into soc-2013-dingto. |
August 17, 2013, 21:24 (GMT) |
Cycles / Sky Model: * Old and new model can now be selected from a drop down menu. * Old blend files load fine and use the old model, when we create a new Sky Texture node it defaults to the new model. * Fixed Strength issues with new model and some code cleanup there. ToDo: * OSL implementation for thee new model. * Code cleanup. Maybe outsource the whole precalc into its own cpp file? |
|