Release v2.5

New Features

• Added the support for Optix/RTX acceleration
• Added the support for Optix denoiser imagepipeline plugin
• GPU imagepipeline now allocated buffers for only the used AOVs by each plugin
• batch.haltspp now supports 2 parameters too in order to have separate halt conditions for eye and light tracing
• Added the support for camera not uniform and anamorphic bokeh (issue #409)
• Added the support for camera bokeh with custom distribution
• Variance clamping is now applied separately to each radiance group (issue #414)
• Added the support for lateral and longitudinal chromatic aberration (issue #409)
• Added the support for SINC pixel filter (issue #415)
• Added the support for CATMULLROM pixel filter (issue #415)
• It is now possible to disable Optix (with property context.cuda.optix.enable) even if it is detected and available
• Any mesh UV map can now be used for displacement (issue #416)
• Subdiv shape now supports multiple vertex UVs, colors, alphas and AOVs (issue #416)
• Added the support for holdout (issue #423)
• Added the support for wireframe texture
• Variance clamping is applied applied to DIRECT_DIFFUSE, DIRECT_GLOSSY, EMISSION, INDIRECT_DIFFUSE, INDIRECT_GLOSSY and INDIRECT_SPECULAR AOVs too
• GetPlatformDesc() now returns compile.LUXRAYS_ENABLE_OPTIX property to check if Optix is available
• Removed the limit of max. 32 AOV types
• Added the support for new AOVs (issue #429): DIRECT_DIFFUSE_REFLECT, DIRECT_DIFFUSE_TRANSMIT, DIRECT_GLOSSY_REFLECT, DIRECT_GLOSSY_TRANSMIT, INDIRECT_DIFFUSE_REFLECT, INDIRECT_DIFFUSE_TRANSMIT, INDIRECT_GLOSSY_REFLECT, INDIRECT_GLOSSY_TRANSMIT, INDIRECT_SPECULAR_REFLECT and INDIRECT_SPECULAR_TRANSMIT
• Added a fast evaluation path in GPU code for constant and image map textures
• Added the support for Distort texture
• Now individual bricks, in brick texture, have a pseudo-random mix between two colors by default
• Added the support for Randomized Texture Tiling (#228)
• Added the support for UVRandomMapping2D
• Added the support for LocalRandomMapping3D
• Added the support for Caustic AOV when using hybrid rendering
• Added the support for TwoSided material
• Reworked WhiteBalance plugin and light groups white balance support to be coherent with usual photographic white balance process (note: this change breaks the compatibility with past scene settings)
• Added a cap (25% and 75%) to Glass transmission/reflection sampling to not under sample some case
• Added the support to all light sources of ".temperature" and ".temperature.normalize" properties to set light temperature
• Added the support for texture bombing
• Added the support for Caustic AOV to OutputSwitcherPlugin
• LuxCore API has new log system
• Added the support for 180 (or any other amount of degrees) view to Environment camera
• Added the support for 180 (horizontal stacked) and 360 degree (vertical stacked) stereo camera (issue #451)

Beta1 released

• Removed the cap to opencl.task.count parameter

RC1 released

Fixed Bugs

• Added the support for pixel filtering to light tracing (issue #413)
• Now LuxCore works even if OpenCL is available but not a valid OpenCL ICD platform has been found (issue #422)
• Fixed a problem with cloth material when using CUDA and Windows (issue #417)
• Fixed a problem in glossy coating albedo evaluation on GPUs
• Fixed a difference between CPU and GPU rendering when using a roughglass with a shadow transparency color
• Fixed the support in .ply files for triangle AOVs
• Fixed a memory leak when rendering with OpenCL (noticeable when rendering animations)
• Fixed text export of brick texture HERRINGBONE and BASKET
• Fixed scale problem in GPU code for brick texture HERRINGBONE and BASKET (issue #438)
• Fixed a possible crash when using trilinear texture with GPUs
• Reworked some condition in hybrid rendering light tracing to solve the problem of flashing caustics (issue #441)
• Fixed a problem with Blender Translation language preference option and scene export on file
• Fixed a self-shadow problem in BIRCPU render engine
• Fixed a Mix material problem when using solid and transparent materials (issue #461)

Beta1 released

• Fixed a crash when using Subdivision shape with vertex colors (issue #469)
• Fixed a possible crash when using multiple GPUs with PhotonGI
• Fixed a deadlock when using multiple renders at the same time in Blender (issue #575)
• Fixed a bug in the support for material .bumpsamplingdistance property
• Fixed a discrepancy in reflected area lights on glossy surfaces between CPU and GPU rendering when light tracing is enabled (issue #470)
• Avoid to return CUDA_ERROR_NO_DEVICE error if CUDA is installed but there are no NVIDIA GPUs installed
• Fixed compilation when CUDA is disabled but OpenCL is still enabled
• Now disabling CUDA for any error during the cuInit() like when an external GPU is unplugged (issue #493)
• Compiling the code with LUXCORE_DISABLE_OIDN defined will disable Intel OIDN (for platforms where it is not available like Apple M1)
• Fixed halt condition rendering stop code when using both eye and light tracing parameters (issue #497)
• Change sphere light implementation to provide an alternative to the point light fallback (issue #443)
• Fixed a compiler error when running with CUDA on Ampere GPUs (issue BlendLuxCore#626)
• Fixed a crash when using an image pipeline with only NOP plugin
• CPU code now rejects NaN/Inf sample values for AOVs like GPU code
• Avoid to generate NaN volume transmittance if the segment is infinite

RC1 released

Check https://wiki.luxcorerender.org/LuxCoreRender_Release_Notes for the full list
of release notes.

Known Problems and limitations

• accelerator.bvh.builder.type = EMBREE_MORTON doesn't work with the latest Embree
• Python crashes when calling sys.exit()
• TILEPATHOCL (and RTPATHOCL) count the layers of transparent surfaces as an iteration of the rendering finite state machine so they require to increase the max. path depth to avoid black regions.
• Light emitted by motion blurred triangle lights is approximated using the triangle area of time 0.0 (computing triangle area at runtime would be expansive) so it may be not totally accurate if the motion blur transformation include a scale.

Note for Windows users: you have to install the Microsoft Visual C++ Redistributable
for Visual Studio 2017 and
Intel C++ redistributable.