I’m surprised there isn’t a thread about this yet!
https://www.solidangle.com/news/press-release-arnold-5/
There are some pretty impressive new features packed in there. And the speed increase is pretty fantastic from what I’ve tested so far. Also, that hair shader is absolutely phenomenal. Would love to see the same controls in Cycles at some point.
Also, their new results for 1 sample/pixel with their improved area light sampling + dithered sampling is just insane:
There is still no Blender integration for Arnold is there? All I have read is that cycles used some SSS techniques from it many moons ago. Why are you surprised?
Because, believe it or not, most 3d artists live outside of the Blender bubble, as evidenced by the countless other non-Blender release threads here. Compounded with the massive popularity of Arnold (even here on BA) I was surprised to see that the usual suspects who trawl the web for new release info hadn’t beaten me to the punch.
What about GPU acceleration in ver.5? After Solidangle hired one of the Cycles developers several years ago, i thought that Arnold will become on par with Cycles much faster.
I’ll be a bit fan boyish here but you can buy a separate render computer for 1095€. Thus making blender really fast.
@Ko.
Ha, you make it sound as if its not on par with Cycles.
They are working on the GPU render, but they won’t release until its as solid as the CPU one, and not without proper marketing.
Autodesk is very silent despite the fact they just released Max 2018 with the new Arnold, which is simply because they have not much to show and tell. If they had this GPU mode finished you would hear it from all sides.
IMHO Arnold will never give support to Blender, Blender is stealing a lot of users to Autodesk right now… and Arnold is owned by Autodesk, me being one of those stolen users hahaha
Cheers.
Is is possible that Solid Angle will no longer publish whitepapers on their newest sampling technology because of the Autodesk acquisition?
One of the neat things about the Solid Angle guys is that they shared some of their techniques through those papers, but they’re likely under a new corporate culture so I’m not sure whether they’ll keep everything proprietary from now on.
Nothing is stopping anyone from Arnold supporting Blender. They have fully-functional and very well-documented Python and C++ APIs. I have fairly basic integration with an older version sitting around here somewhere, and a more talented coder than myself could probably have something functional up and running with the newest version in a few days.
Arnold is way ahead of Cycles in terms of the total feature-set though (there is a reason why it’s become one of the most heavily used engines in Hollywood).
Outside of GPU rendering, Arnold is the best engine out there that’s actually available for anyone to purchase (as even more advanced engines like Manuka are in-house solutions only), though Renderman is looking to be a serious competitor very soon if it’s not already.
No one in Hollywood is ready to really get into it and adopt GPU rendering in production, especially from an off-the-shelf rendering solution. The limitations are still far too rigid for real-world use outside of previewing and pick up shots. When the technology is there, trust that all of the big names in the industry will be quick to adopt a GPU-based pipeline. But for the time being, the reliability and existing pipeline of CPU farms aren’t worth throwing away for an architecture that still gets completely overhauled on a yearly basis.
Fair enough, I was hoping that it was somehow linked to Blender in a way that I didn’t know about.
Two. Brecht (creator of cycles) and Mike Farnsworth (HDR support and MIS for HDR lighting)
Awesome stuff, thanks for the update @m9105826. Didn’t hear anything about a new release yet anywhere else. I’ve only recently been able to experiment more with a Blender-to-Arnold pipeline by making use of Image Engine’s Gaffer, and I’ve been sold on it.
One of the big selling points of Arnold and RenderMan is that they can handle an insane amount of geometry and texture data. And that is exactly where most GPU-based renderers fall short: everything is fine there as long as your scene fits into VRAM. As soon as it exceeds VRAM you fall of the performance cliff (as ChaosGroup’s Vlado put it) and suddenly your render is only 10% as fast as before. It’s not like production didn’t look at GPUs - Weta was using their own GPU renderer (https://research.nvidia.com/publication/pantaray-fast-ray-traced-occlusion-caching-massive-scenes) before Octane was out.
Another limitation of GPU rendering is limited support for programmable shading. Traditionally, RenderMan shaders have been used and abused to do the impossible - it was used to ray trace inside RenderMan before RenderMan could ray trace, it was used for debugging output, file format support, on-demand geometry creation, etc. Likewise, shaders in Arnold can be used to build new features into it that it didn’t have before. A lot of this is either not possible in GPU rendering at all or only at abysmal performance.
All I have read is that cycles used some SSS techniques from it many moons ago.
I am not so sure about this, maybe someone else knows for sure. I don’t think it [cycles] uses straight up tech from Arnold for the SSS shader. What I’ve seen though is someone creating a SSS Skin shader with Cycles nodes that mimics the skin shader in Arnold.
But straight up using tech from the renderer, mmmmm. I dunno.
This is the technique published by solid angle that’s implemented in Cycles:
Again GPU vs CPU, many people ignore one very important fact: Big companies dont care about money, and with extremely fast Xeons and now als the AMD Maple Architecture (32 Cores!) GPUs are getting behind flexibility/price (not performance/price) ratio.
Not to mention if you have 32K multilayer textures like in Gravity.
Now, not so fast. The future of GPUs is (still) just on the rise. Performance wise they ARE the future. The problem is if you want to make a huge renderfarm that NEEDS I/O coherency. We are not there, yet. Vega and Volta will change that. There were already Vega prototypes at CES with an 1 TERAbyte (yes you heard it right) SSD, beside 16GB HBM VRAM on GPU! They showed off 250 GB Texture PT in realtime!
So it come, but the big renderers will not jump on the GPU wagon UNLESS you have the same Memory (nUMA or UMA cluster) and FAST interfrabrics from GPU to GPU. Both AMD and NVIDIA are still infants regarding that now.
Never forget, a renderfarm is high-performance computing. And they dont care about money (primarily), but “can we 500 to 5.000 people at this or that VFX/3D company do that?” 
Sorry for OT, still no Gaffer for Windows?
Anyone tried with the exporter ie. btoa from noxman or similar?
Not sure if it’s been posted anywhere else, but the newest release has been announced. Looks like they settled on Optix as their backbone for GPU rendering, marking yet another production renderer whose developers who have decided that OpenCL isn’t worth the headache.
