OctaneRender™ Blender 3.3 plugin version 2021.1.6 - 26.5 [STABLE] released

This link send you to Pro users page… you can download it, but you can’t use it. I once make such mistake :laughing:

What do you mean? It lets you download Blender_Octane_Edition_2021.1.4_26.0_beta right? What happens after you install both the program and octane server?

If you install payed version and try to render you get error message. First picture show how it look when you try to download payed version.
Second picture show how look download page for free tier… Octane which work.
Well, maybe I’m wrong, try it and show result. Maybe I do something wrong.
Octane check you license status, if your account are limited to free tier, he don’t work.

Free tier is available since April 1st.

You just need to scroll further down on the link.

Side note, anyone find Octane initializing particle hair system much slower than Cycles?

1 Like

Ok thanks a lot
We are not proposing octane shaders for our starter library, waiting for official 3.0 support to support the asset browser properly

I must say, currently on the go with a laptop & 4gb of VRAM and octane was very hard to use :melting_face: Couldn’t load more than a few vegetation assets w/o going out of memory. Any tips to clean up VRAM while using blender?

For a laptop it’s going to be difficult, also 4gb is about as low as you can go with Octane. Make sure that you don’t have any virtual desktops in use. I would shutdown any other programs that are running as well although it will probably not make much difference. Turn on out of core memory, probably not help with instances, but it will help with textures and geometry. Having more system memory will help with this as you will be using it instead of vram (see shutting down other programs).

This would be on the more extreme side, but you could install Linux on your laptop, if you are not already running it. Linux’s footprint is much smaller than Windows, uses resources more efficiently, and can be a bit faster in some cases. This probably will not affect vram usage much as it would be the same with Octane for either OS.

There are some other things you can do within Octane like choosing a render kernel that has a smaller footprint on the GPU (not sure which one would be the best, my guess would be Direct Light). Some settings in Octane will use less vram than others as well, but that is getting into the weeds and I’m not sure how much vram it would save you. It would depend on the scene, but basically any setting that sacrifices vram for speed would be the ones I would concentrate on.

Edited to add…
Run an older version of Octane, the older kernels use less vram than the new versions, unless you need some functionality that would be missing in an older version.

1 Like

Didn’t knew it was possible thanks !

My main problem is chrome browser, it’s using bit of GPU memory too I believe


Thanks for helping out !

1 Like

Just noticed that Simplify in Octane doesn’t have a texture limit. Is there another way to texture limit?

Hi all

We have made many improvements since XB0 and we’re happy to bring you XB1 of 2022.1.

We’re also releasing XB1 for MacOS, but as mentioned previously, please be aware that there are some known render regressions we are working on. Also, at the moment it is not possible to use CUDA render nodes with MacOS standalone.

The MacOS builds below will only work correctly on MacOS12 and Apple GPUs.

As with all experimental builds, we advise to not use this release for production purposes. We are trying to, but we can’t guarantee that scenes saved with this version will be compatible in future releases.

Changes since 2022.1 XB 0

  • Added an option to toggle bump/normal mapping for the ambient occlusion AOV.

  • Compositing for output AOVs will now happen on the GPU - this is much faster but uses more GPU memory and can be turned off in device preferences.

  • Added more to the Procedural effects texture node:

    • Blaschke product
    • Fire emitter
    • Fractal 3
    • Mist
    • Skinner
    • Sun surface
    • Tunnel
  • Added unary Vectron operators, Clip and Offset.

  • Added binary Vectron operators Ink, Inset, and Intersect.

  • Added Bounds, Radius and Transfer material to Vectron Subtract node.

  • Added Bounds to Vectron Union node.

  • Added Vectron domain operator node.

  • Added a compression rate setting to the batch render job.

  • Added new texture nodes to generate fields: Angular field, Planar field, Shape field, Spherical field

  • Added Vectron primitives: Box, Capsule, Cylinder, Prism, Sphere, Torus, Tube

  • Added a “Unit volume” geometry node

  • Added additional Length and Normalize vector operators to the Unary math operation texture node

  • Added utility nodes for float, int and bool primitive values in the Values category: Float to int, Int to float, Lookat to transform, Boolean logic operator, Float relational operator, Int relational operator, Binary math operation, Unary math operation, Range, Rotate, Float merger, Float component picker, Float if, Int merger, Int component picker, Int if

  • Updated the volume import preferences to support loading of grids that are not assigned to a specific use case (like scatter, absorption, etc.) yet. Standard volume mediums can select these grids by name on a per-medium-instance basis.

  • Lua API: in saveImage3, saveRenderPass3, saveRenderPasses3 and saveRenderPassesMultiExr3 the exrCompressionType argument is replaced with
    the formatOptions argument which accepts a table with options. A numeric value is still interpreted as a compression type.

  • Added a “Legacy blend mode” alpha operation and a force pre-multiply opacity pin to the composite AOV layer node.

  • Added Instance position projection node.

  • Added a “no shading” option to clipping material.

  • Image nodes now have a Border mode for U and V separately.

  • Volume to texture node can now read 3-channel VDB grids.

  • The Read VDB node can now access VDB grids by name.

  • Prevent the power input of a texture (when it is used as a normal map) being evaluated as a normal.

  • Version conversion to the new single scatter factor input on medium nodes is more accurate.

  • Modified the Standard surface SSS implementation to avoid the confusion in forward anisotropy changing the observed albedo.

  • Added light linking to toon shading for toon lights.

  • Implemented single scatter events in volumes.

  • Removed legacy emission pin from emission nodes and goes back to their old behavior.

  • Tangent normal now always shows the shading normal with bump/normal mapping applied.

  • Added diffuse transmission thin wall option to universal material.

  • Improved performance of hair rendering, which is now faster and uses less memory, and behaves correctly when animated.

  • Added real-time mode (lightning button on render viewport). In this mode, if possible, the film buffer is never moved between the host and the device. This means that compositing has to be done on the GPU, and it also enables a “real-time” display, which will display rendered images faster. These features require more VRAM to be available. It does not have a large effect on MacOS because Apple GPUs have Unified Memory.

  • Added ability to choose premultiplied alpha for EXR files saved by batch rendering script or batch render job.

  • Added voxel interpolation settings to Standard volume medium including nearest-neighbor, trilinear and tricubic.

  • The Mesh volume node now works with Standard volume mediums.

  • Standard volume medium channels now support color data (except for the temperature channel).

  • Added support for the ‘Channel’ emission mode to Standard volume medium.

  • Experimental support for photon mapping in medium and volumes.

  • Added native ACES tone mapping without the need for OCIO.

  • Increase maximum photon depth to 256 for pure specular materials

  • Cinema4D noises are now available in all plugins.

  • Increased the maximum framerate of various animation scripts to 1000 fps.

  • Improved spacing of tonemap intervals near start of rendering.

  • Added fake shadow and refract alpha to Standard surface material.

  • Enabled the denoiser for the Photon tracing kernel.

  • Fixed possible crash while loading a new project.

  • Fixed scene file color space being ignored sometimes.

  • Fixed random walk SSS seams when using nested dielectrics.

  • Fixed a crash when trying to use the Edit menu when using Octane UI via a plugin.

  • Fixed an issue where a material with subsurface scattering and fake shadows enabled renders too dark, when it is inside another material with fake shadows enabled.

  • Fixed render hang when using the baking camera with displacement triangles.

  • Fixed a crash while opening certain scenes that were created in an older version.

  • Fixed a crash when deleting an input to a Vectron node.

  • Fixed crash when loading Vectron sample scene.

  • Fixed Vectron rendering issues when size pin is disconnected.

  • Fixed handling of transforms when using Vectron transforms in a texture shader.

  • Fixed transforms sometimes being applied more than once to a Vectron node.

  • Fixed an issue when copy-pasting a geometric primitive.

  • Map range output AOV node will now allow the output minimum value to be higher than the output maximum value.

  • Fixed transparent border around render region when using the info channels kernel.

  • Fix artefacts in Z-depth front pass where multiple volumes are close to each other.

  • Fixed missing tooltips in windows created by plugins.

  • Fixed the node search box appearing behind the node graph editor window when created by the plugin.

  • Fixed rendering black pixels when light opacity or general opacity is 0.

  • Fixed universal material transmission type not updating correctly for back faces if material layers are used.

  • Fixed universal material too bright when metallic green or blue channels are above 0.

  • Track frame rate and optionally show it in the render viewport status bar.

  • Fixed a crash when you delete a volumetric light that is connected to a render target, when rendering has already started.

  • Fixed incorrect starting frames when using batch rendering.

  • Fixed clipping material casting a shadow when light general visibility is not 1.

  • Fixed standard surface subsurface layer (without specular) breaking when used on multiple meshes.

  • Fixed accurate rounded edges rendering incorrectly on Turing devices.

  • Fixed glossy material inside glass rendering incorrectly with nested dielectrics.

  • Fixed motion blur for USD geometry.

  • Fixed USD animation not working on a copied USD geometry archive.

  • Fixed a crash loading USD scenes with curves.

  • Fixed unnecessary geometry re-processing times when switching between universal camera nodes, if you use vertex displacement and the texture has inputs outside the root of the texture node tree.

  • Fixed vertex displacement on Ampere devices.

  • Fixed smooth terminator not working with multiple toon, quad, or sphere lights.

  • Fixed net rendering with AI light update causing tile seam artefacts.

  • Fixed crash when a material/object layer map node has a scatter on surface as input.

  • Fixed crash of net render nodes if the scene uses MLT and info AOVs.

  • Fixed broken render passes export UI after previously saving deep EXR when deep image is no longer enabled.

  • Fixed incorrect subsampling of shadow passes.

  • Fixed a composite AOV bug where changing layer opacity caused incorrect results.

  • Fixed render jobs only working once sometimes.

  • Fixed scene updates when changes are made with Cryptomatte AOV being tonemapped in the background.

  • Fixed artefacts in cryptomatte passes if using net render and render region.

  • Fixed crash when a user instance ID cryptomatte AOV is being saved.

  • Fixed a missing pin when loading an Alembic file.

  • Fixed blend mode alpha operation of composite output AOV node which was causing blending operations to produce an incorrect result.

  • Fixed AI light update making scenes heavy on emitting primitives very slow.

  • Fixed render failure when using too many scatter on surface.

  • Fixed white square in the viewport on Linux in some situations.

  • Fixed light linking through transparency surfaces giving unexpected results.

  • Fixed FBX geometry normals.

  • Fixed render failure when using a texture with Sample pos. to UV projection for vertex displacement.

  • Fixed render failure when using clay mode and mix material together.

  • Fixed incorrect focus on input linker when “Show in graph editor” is selected for a material node that is connected to a scene node graph.

  • Fixed emission with “visible on diffuse” disabled not working with universal material coating.

  • Fixed loading of FBX images with relative paths.

  • Fixed skeletal animation not working if a joint node is not animated.

  • Fixed net render crash caused by AI light.

  • Fixed image exposure settings being ignored on image and color output AOVs.

  • OSL: Fixed value for P if a Vectron node is created from a Volume to texture node.

  • OSL: Fix possible crash when using a constructor to assign a struct variable

  • Add an input for photon rays to the ray switch node, and support raytype(“photon”) in OSL

  • OSL: textures can now create texture inputs that are evaluated as normals. This is achieved by declaring the input as a normal in the OSL shader, and adding the [[ string widget = "color"]] metadata.


OctaneRender™ for Blender plugin. Windows, and Linux versions. The download links, provided on this page, are for Enterprise License Holders.

OctaneRender Studio subscribers can access their software from the “Downloads” section of their Customer Account Page.


  • Upgraded to the Blender 3.1.2


  • Fixed the Denoiser render issue in the final render mode
  • Fixed the crash when using Cycles
  • Various Bug fixes

When will the free tier version will be updated to match this?

It appears it has been.
Haven’t downloaded it yet to check.


OctaneServer Prime (2021.1.4) 26.1 stable:

Blender Octane Edition (2021.1.4) 26.1 stable:


Weirdly the MacOS Metal version is still stuck at Blender 2.83. I wonder if they will update it at all?

That would require OctaneX and while its already there for C4D, there are still many issues.I guess the plugin developer wait for support before adding it. Kind of Otoys problem, to many external developers. Redshift is progressing slower, but in higher level and the Blender plugin is one of the few external render, that are true add-ons.

So is Octane finally an addon or is it still a separate build?

1 Like

Still a separate build right now. :frowning:

1 Like

Rumors are we won’t see an addon version until Octane 2022 version.

1 Like

Separate build is not big issue for me. After all, always can save scene in ‘normal’ Blender, and open in Octane Blender. I meantime can take beer or two :wink:
I wish that there is some even most simplified texture preview in viewport for texture. When import image texture and adjust scaling, projection, it’s inconvenient to have GPU rendering all time. But I guess that this is not possible, Octane for C4D also don’t have such possibility, for example.
btw. this can be interesting comparison, great artist which render same scene in C4D with Octane and Redshift ( hardcoding scene ).
ZAKI ’ on Behance
So chose your poison :slight_smile:

Well, it’s not their fault. The Blender SDK is said to be in a bad shape and many company and add on developer complain of lack of support for C++ Api and more control for add-ons to manipulate the host application.
So far, its for them twice the work. And more then this plugin render work on how shader nodes work in other application, just not like it is in Blender.
On order to bring things together both parties need to move. And I think the developers of Octane have a huge knowledge of how to add their stuff. Its just Blender Devs seem to take the whole thing not as important.
For example, if I use Octane in Modo, everything is supported without even adding a single thing to Octane. So all textures and shaders work. Special functions in shader can be obtained by adding an overwrite, that automatically translate the shader to an Octane shader. So I still can use the default render and still have the Octane override. Both live in the same world.
Of course going deeper you need to know much more and test things. Like sport likes being different, displacement and noises work different etc. But if I take a normal Modo scene and just let Octane render, I get almost the expected result, without changing anything. Blender is lightyears away from that.
Things might get better when Blender supports new shader formats that allow unified shader definition. Maybe that can make even Blender shader work in other renders… maybe, who knows.