Blackmagic Forum

Greetings,

Having failed to find a current/comprehensive roadmap for planned Fusion (and Resolve/Fusion) development, I'd like to just go on the record here: It seems high time that the 3D rendering functions were upgraded to include a physically-based material/lighting model; Or, in place of improved internal tools, Fusion should allow for plugging-in of 3rd party renderers.

Whether standalone or under Resolve, Fusion's power as VFX tool would appear to be in its tight integration of 2D and 3D functions (bravo!). But, though it's very nice to be able to bring in scenes via FBX import (bravo again!), it is not so great that they appear to have been rendered in the mid 90's.

Sorry for the gripe, but Fusion's 3D rendering is starting to look increasingly old-school and in need of modernization. Is there any evolution in the works?

Thanks.

m

tekkonkinkreet wrote:Fusion should allow for plugging-in of 3rd party renderers.

That specific part of your request is already present and has existed in Fusion Studio since June 2012.

Fusion Studio supports 3rd party renderer integrations via a, free to license from BMD, C++ based FusionSDK library. It integrates the external renderer as a pull-down option inside the standard Renderer3D node. The capacity has existed all the way back to pre-BMD times with eyeon Software's Fusion Studio v6.4 release.

Fusion customers over the years at VFX companies have compiled a PIXAR RenderMan (pre-RIS era) plugin integration for Fusion, along with a 3Delight Studio plugin renderer (with the full source code posted here), a Disney Parti I/O plugin, and even a ThinkBox Kratatoa based particle plugin rendering solution was possible at one time.

There are even after-market plugin shader systems like Chad Capeland's extremely powerful CustomShader 3D for Fusion 9 on Windows, and AccuShader's Disney principled BRDF based PBR materials for Resolve/Fusion v17.

The main issue as I understand it, is that there is a user-base "size" problem for a 3rd party company to effectively dedicate the time and resources to make and provide ongoing support for a full-featured renderer integration that would exist inside a compositing toolset.

And there is the issue that casual, extremely light-duty users of a potential Fusion based 3rd party renderer plugin, might alternatively be able to solve 80% of their needs, at zero cost, by opening up a copy of Blender and using Cycles or Evee on a 2nd monitor on their PC, with a Fusion Studio comp open on the primary display.

There's also the issue of BMD's lack of public commitment to a publicly documented SDK for developing plugins as well as the licensing restriction that 3rd party plugins can only be used with Fusion Studio.

Those issues aren't exclusive to 3D rendering plugins, but there's also no way around them. You can't use OFX or ResolveFX or DCTL for 3D plugins.

Thank you, Andrew, for your kind and detailed reply. While I was waiting for my first post to be approved, I stumbled on some other things: Reactor, which has your fingerprints all over it (bravo), as well as the ancient 3Delight plug project you mention (wish that was being maintained).

Andrew Hazelden wrote:The main issue as I understand it, is that there is a user-base "size" problem for a 3rd party company to effectively dedicate the time and resources to make and provide ongoing support for a full-featured renderer integration that would exist inside a compositing toolset.

Yes, that seems the most likely explanation for the scarcity of such 3rd-party tools; and, in my opinion, all the more reason to push for improvements to the on-board rendering.

The direction we are moving here with our use of Fusion (and Resolve), we tend to lean hard on the 3d side of the toolset. (That would seem to be the direction of many users who would rely on Fusion as their main VFX tool.) And the time lost moving data back and forth to and from various DCCs is not inconsequential (FBX and Alembic files always seeming to need a massage one way or the other).

Having to periodically finish shots on our single Nuke seat when we hit a wall with Fusion’s 3D is a drag. (Having to finish in AfterEffects now and again is kind of a downer too, though not specifically related to Fusion’s 3D functionality.)

Thanks again!

m

tekkonkinkreet wrote:Having to periodically finish shots on our single Nuke seat when we hit a wall with Fusion’s 3D is a drag. (Having to finish in AfterEffects now and again is kind of a downer too, though not specifically related to Fusion’s 3D functionality.)

As with most things in life, there are many interesting workrounds / haxies that are possible with Fusion to bolt on better rendering tech that can be shoe-horned into a comp, but they don't fit in with the "out of the box" nicely integrated 3D renderer goals people have.

Kartaverse X V-Ray X Domemaster3D Lens Shaders
https://www.linkedin.com/feed/update/ur ... 099454464/

tekkonkinkreet wrote:Yes, that seems the most likely explanation for the scarcity of such 3rd-party tools; and, in my opinion, all the more reason to push for improvements to the on-board rendering.

Thanks again!

m

To be blunt.. what makes anyone think if BMD can't support its existing tools and help 3rd party devs that somehow BMD will somehow pick up the torch, after years of decay and start updating its 3D tools again... or any tools for that matter? We can throw all the love and hope at Fusion but historically it's been pretty dismal for years with regards to development. If it weren't for Reactor and the awesome people supporting it, and the lower cost of Fusion, it would be dead. Although the lower cost and perpetual licensing (in this case upgrades forever) hampers any extra ability BMD would have to fund extra development in Fusion. This is just the way it is with Fusion. As others mentioned.. gotta do hacks, workarounds or render in your favorite DCC instead.

I will say it's great to hear about Unreal Engine in Nuke, although I've yet to play with it. Virtual sets and production just got a boost.

Jeff Ha wrote:To be blunt.. what makes anyone think if BMD can't support its existing tools and help 3rd party devs that somehow BMD will somehow pick up the torch, after years of decay and start updating its 3D tools again... or any tools for that matter?

I've had a lot of alternating views over the years on how Fusion has evolved as a standalone tool and inside of Resolve's Fusion page.

What one defines as "decay" may vary if you change from talking about a bit of stagnation on the nodes in Fusion (like a lack of innovation in DeepEXR, PBR Metal Roughness, 360VR, image based modelling, better multi-channel support, etc) vs the core-engine inside Fusion that works cross-platform.

On the development side the Fusion team at BMD should be commended for the effort they all put in to make Fusion Studio and Resolve able to run on Apple M1 (Arm) CPUs as pretty much a launch day customer supported thing. This was no easy feat and required a major re-think, and under the hood change of a lot of features long-time customers might not have noticed at all.

Jeff Ha wrote:To be blunt.. what makes anyone think if BMD can't support its existing tools and help 3rd party devs that somehow BMD will somehow pick up the torch, after years of decay and start updating its 3D tools again... or any tools for that matter? We can throw all the love and hope at Fusion but historically it's been pretty dismal for years with regards to development.

Ooh, that’s really not what I want to be hearing.

I’ve been using Resolve to edit, grade, and do the odd bit of touch-up, now for a couple years, and have really appreciated its workflow and performance, across platforms and configurations; so much so that recently I have been trying to move my team over to Fusion as a core VFX tool. Maybe I was too optimistic in imagining that Fusion (embedded or standalone, I really don’t care which) could be that tool. But the users’ opinions on display here have begun to convince me that it is my perception of Fusion that needs upgrading. Perhaps I should think of Fusion as less of a potential core VFX tool, and more as a sidecar -- a utility for accomplishing minor VFX pick-up tasks within the editorial context (i.e., the way I had been using it).

I appreciate being able to develop homegrown solutions to suit our unique situation here, in Japan, creating hybrid (stop motion / hand-drawn / CG) films. Strong SDK and interoperability tools are a must. But I also don’t want to be forced to switch softwares or go full DIY every time an onboard tool gets me only part of the way there. (I realize I am speaking in generalizations, but several times now we have been beguiled by features that promise much, but don’t quite deliver the goods in the home stretch.) Esoteric tasks demand more esoteric solutions (of course). But there are some fairly pedestrian bits of Fusion that are seriously undercooked (3D being my main complaint). And invariably it is quicker to switch softwares than to attempt to code or script our way out of those problems when in the context of an ongoing production. Quicker, yes, but costly nonetheless. I had imagined Resolve and Fusion as an afordable alternative to the vastly overpriced industry standard (Nuke) and the Adobe ecosystem with its steamer trunk full of plugins that are de rigueur for AE users. But now I am discovering that (worst case) just to finish tasks I'd thought I could tackle with the BMD tools alone, I also need seats of both of those other tools (and that's in addition to the 3D DCC) to switch over to when Fusion's -- ahem -- spartan 3D implementation is found lacking.

I think the current entry point (under ¥40,000 here) for the pro version of the software would still be very attractive for users here in Japan, even with periodic maintenance or upgrade fees (can't believe I'm saying this!), if that is indeed what is needed to guarantee some forward momentum in R&D. And I think many of us wouldn't mind periodically paying a bit more in exchange for innovation and ongoing development of advanced features. That's probably not a popular point of view here, and another topic entirely.

Best,

m