Hi all
I’m researching a transition at my shop from Nuke and Flame to Fusion, for basic/intermediate Composition And Motion Design applications. Would be nice to be all BM centered. I guess basic questions are:
1) the learning curve/transition - I figure we would a masterclass/support for the beginning. Basically how bothered you recon a experienced flame or nuke artist will get with this kind of transition? Of course is a subjective question, but how far away are this softwares operation from each other?
2) can fusion make everything that Nuke and flame can (again, for basic/intermediate comps and motion)? Is it as fast and as reliable?
3) I understand that adding a Mocha is paramount. Would it suffice for tracking? Are there any other add ons to help it out?
4) Stand-alone version is a must? How nice does it interact with Resolve?

Thanks a lot


Enviado do meu iPhone usando Tapatalk

Hi. If you are looking for a masterclass that covers training on Fusion Standalone and the Fusion Page in Resolve for your team, I’d like to recommend either Eric Westphal or Sander de Regt.

They are both experts in their fields and excellent communicators.

https://de.linkedin.com/in/ewestphal

http://fusiontrainer.de/

https://nl.linkedin.com/in/sanderderegt

Also, you should check out the Steakunderwater Fusion Community site as well. It’s loaded chock-a-block full of knowledge and useful content.

1) In my experience, a Nuke artist transitions to Fusion quite quickly, especially if the Nuke2Fusion bundle from Reactor is installed—this is a set of hotkeys and macros that make the Fusion environment much more familiar.

I've trained several Nuke users, and they can be doing general comp tasks within hours. Maybe two weeks for a mid-level artist to become about as proficient in Fusion as they were before. I've never assisted a Flame user's transition, but we did have two Flame artists who joined our company at the same time I did, and they didn't seem to have any trouble (I was transitioning from Nuke at the same time).

Someone coming from Nuke might struggle at first with the channels paradigm—Fusion doesn't have anything like Nuke's arbitrary channel system that you can pass down the pipe—you wind up with more parallel wires to bring mattes and so forth downstream. And people who have done a lot of projection or other 3d work might find Fusion's 3d system constraining. I never got very deep into Nuke 3D, though, so I'm not sure where the pinch points will be.

2) I've never tested both programs on similar materials side-by-side, so I can't quantify the speed differences, but Fusion has always felt faster to me. Maybe that's because it doesn't tease you with that scanline slowly drawing the image—you get the whole frame at once. Or maybe it actually is faster; I don't know. Stability seems about equal, although that can depend greatly on configuration and how you use the tools. I've had a lot fewer licensing headaches with Fusion. Plug in the dongle; done.

Capabilities are broadly similar. Fusion doesn't have Weta-style Deep compositing tools, and there are some plug-ins available for Nuke that aren't offered for Fusion (KeenTools, for instance). I've not heard of much that Nuke can do that Fusion cannot. Any it's easy enough for a competent technical artist to reverse-engineer most Blinkscripts or Gizmos into Fuses or Macros. I've done that a couple of times, and I know there's somebody else working on converting the most popular gizmos.

3) Mocha is magic. No question. I would recommend that you also plan on eventually adding a dedicated 3d tracker, also. Syntheyes is affordable. I use PFTrack, but it's pricier. Fusion's CameraTracker might suit your needs for a while, but it definitely has its limitations. Both Syntheyes and Mocha can export Fusion nodes without the need for additional plug-ins or scripts. I think the folks who make Silhouette were also working on a Fusion export module. I'm not sure if anything came of that.

4) From your questions, I'd say yes, the Standalone is a must. The Fusion page integrated with Resolve can do some cool stuff, but for visual effects tasks, it's better to run without Resolve's significant resource overhead. And comps in Resolve's Fusion page are locked up in the database instead of being discrete files on disk, so they're both less robust and less flexible.

I haven't tried to build a Resolve -> Fusion pipeline yet, so I couldn't say how nice the interchange is. I'm sure there are others out there who could tell you more on that score.

And finally, I second the recommendation for both Sander and Eric—they're both extremely knowledgeable and good teachers.

Andrew Hazelden wrote:Hi. If you are looking for a masterclass that covers training on Fusion Standalone and the Fusion Page in Resolve for your team, I’d like to recommend either Eric Westphal or Sander de Reget.

They are both experts in their fields and excellent communicators.

https://de.linkedin.com/in/ewestphal

http://fusiontrainer.de/

https://nl.linkedin.com/in/sanderderegt

Also, you should check out the Steakunderwater Fusion Community site as well. It’s loaded chock-a-block full of knowledge and useful content.

Thanks Andrews
I’ll check your recommendations right away.


Enviado do meu iPhone usando Tapatalk

Bryan Ray wrote:1) In my experience, a Nuke artist transitions to Fusion quite quickly, especially if the Nuke2Fusion bundle from Reactor is installed—this is a set of hotkeys and macros that make the Fusion environment much more familiar.

I've trained several Nuke users, and they can be doing general comp tasks within hours. Maybe two weeks for a mid-level artist to become about as proficient in Fusion as they were before. I've never assisted a Flame user's transition, but we did have two Flame artists who joined our company at the same time I did, and they didn't seem to have any trouble (I was transitioning from Nuke at the same time).

Someone coming from Nuke might struggle at first with the channels paradigm—Fusion doesn't have anything like Nuke's arbitrary channel system that you can pass down the pipe—you wind up with more parallel wires to bring mattes and so forth downstream. And people who have done a lot of projection or other 3d work might find Fusion's 3d system constraining. I never got very deep into Nuke 3D, though, so I'm not sure where the pinch points will be.

2) I've never tested both programs on similar materials side-by-side, so I can't quantify the speed differences, but Fusion has always felt faster to me. Maybe that's because it doesn't tease you with that scanline slowly drawing the image—you get the whole frame at once. Or maybe it actually is faster; I don't know. Stability seems about equal, although that can depend greatly on configuration and how you use the tools. I've had a lot fewer licensing headaches with Fusion. Plug in the dongle; done.

Capabilities are broadly similar. Fusion doesn't have Weta-style Deep compositing tools, and there are some plug-ins available for Nuke that aren't offered for Fusion (KeenTools, for instance). I've not heard of much that Nuke can do that Fusion cannot. Any it's easy enough for a competent technical artist to reverse-engineer most Blinkscripts or Gizmos into Fuses or Macros. I've done that a couple of times, and I know there's somebody else working on converting the most popular gizmos.

3) Mocha is magic. No question. I would recommend that you also plan on eventually adding a dedicated 3d tracker, also. Syntheyes is affordable. I use PFTrack, but it's pricier. Fusion's CameraTracker might suit your needs for a while, but it definitely has its limitations. Both Syntheyes and Mocha can export Fusion nodes without the need for additional plug-ins or scripts. I think the folks who make Silhouette were also working on a Fusion export module. I'm not sure if anything came of that.

4) From your questions, I'd say yes, the Standalone is a must. The Fusion page integrated with Resolve can do some cool stuff, but for visual effects tasks, it's better to run without Resolve's significant resource overhead. And comps in Resolve's Fusion page are locked up in the database instead of being discrete files on disk, so they're both less robust and less flexible.

I haven't tried to build a Resolve -> Fusion pipeline yet, so I couldn't say how nice the interchange is. I'm sure there are others out there who could tell you more on that score.

And finally, I second the recommendation for both Sander and Eric—they're both extremely knowledgeable and good teachers.

Hi Bryan
Thank you very much for your detailed explanation. It really clarified a lot of things for me. Best, Marco


Enviado do meu iPhone usando Tapatalk já

Bryan Ray wrote:Someone coming from Nuke might struggle at first with the channels paradigm—Fusion doesn't have anything like Nuke's arbitrary channel system that you can pass down the pipe—you wind up with more parallel wires to bring mattes and so forth downstream.

Can Wireless Link nodes fill in for this somewhat? I saw a YouTube (Pirates of Confusion or something like that) where the guy had a HUGE Fusion layout with strips separating about 7 columns wide and rows for each layer -- each "cell" having a bunch of nodes. He had most of the Loaders in one column and then used wireless nodes to communicate them to the other cells. (Haven't used those nodes myself.)

Sure, but it still adds complexity. The Nuker wants to have a single Loader containing a multi-part EXR from which they can shuffle out channels at need. Lacking that capability, Fusion requires multiple Loaders, even if they're all looking at the same file, and more attendant nodes to manage them. It can, of course, be mitigated with automation and things like the Wireless Link, but it's never anywhere near as elegant as Nuke's way of working.

Which is not to say that I want to go back to Nuke any time soon; I really like Fusion, and in spite of BMD's missteps with the product, I still like them better than Foundry (or more to the point, Foundry's holding company).

Agreed. I actually did a bit of work with Shake back in the day, and then Nuke. I remember the ease of making new channels that you could tap into farther downstream without explicit wiring. Though I never really used that, so don't feel the pain of an experienced and sophisticated Nuker.

Personally, Fusion Studio is one of my most enjoyable programs. I'm not so sophisticated that I bang up against its limitations, it's light-weight (in a good way) and reminds me a lot of Nuke. (Nuke being a distant memory now for me.) I use Resolve more intensely, but Fusion is just plain fun to use.