Fusion vs NUKE

What do you say ? Nuke or Fusion ? Atm i would say NUKE is more used by the Big Post Pro Companys.

Even Nuke is now so much more expensive i think they will change there Price soon. They need to.

But Nuke seems to be better in some cases. What do you say ?

Mike

Nuke is more advanced than Fusion right now. BMD may change that. If you look at what BMD has done with Resolve, it wouldn't be unreasonable to expect Fusion to go through many updates quickly and rapidly improve.

Also, Nuke is industrial-level expensive. They have a free version coming out, but it's for non-commercial use only, so it won't solve any business problems.

Nuke also integrates with Mari, Foundry's 3D paint program. This becomes really useful for matte paintings, among other things. BMD is going to need to address that if they want to compete with Nuke.

Until Fusion seeps into the VFX community, Nuke training will be a more valuable skill. If you were starting VFX for your own productions right now, Fusion is a pretty sweet deal.

It's a bit of a silly topic, but given the current circumstances I'm itching to play along.

This is a testament to how much Eyeon's marketing sucked in the past.

The Nuke-Mari bridge consists of nothing that Fusion can't do with some scripts added. If there's enough demand, BMD or even some users here can get that working in a very short time.

It's true that Nuke does certain things very well, and arguably better than Fusion. In some cases a LOT better (please read Stefan's post about linear workflow). Calling it therefore more advanced, though, no. I would argue that Fusion is - at it's core - a much more advanced tool than Nuke.

Please tell me how Nuke is resolution independent, for example.

I heard somewhere that NUKE doesn't utilize GPU for rendering while Fusion does. If that is indeed true, that's a huge advantage for Fusion in my book (especially since Aftereffects doesn't utilize the GPU either).

Anyone have any info on that?

The 'which software is better' questions never go away, do they? Nuke and Fusion are both excellent products. It's the artist that makes the final images look great, no matter what software they use.

Atm i would say NUKE is more used by the Big Post Pro Companys.

Yes, this is true. It also doesn't make Fusion less of a product. The reason that studios choose particular software isn't necessarily due to what is 'better'. Marketing is a major part of who buys what. Outside of the 'big post companies', you'll find After Effects used A LOT. There are also several companies using Fusion (including the one that I am at right now).

All of them are great tools, but that is all they are: tools. The artist that uses the tool is what is important.

I heard somewhere that NUKE doesn't utilize GPU for rendering while Fusion does. If that is indeed true, that's a huge advantage for Fusion in my book (especially since Aftereffects doesn't utilize the GPU either).

Anyone have any info on that?

Fusion can use OpenCL on some node to speed up processing, but nuke also has some nodes with OpenCL/Cuda code.

Fusion can use OpenGL to render the scenes from the 3d space

I use Nuke and I like it, I Assume that fusion is not much different.......

Choice? Nuke. (for no particular reason but because I know it...)

Having used both in production here is my quick comparison list.

Fusion Pros (vs Nuke)
-Overall faster rendering
-More intuitive
-Better 3d toolset
-Superior text/titling tools
-PRICE!!!

Cons
-Clunky management of linear color workflow
-Poor support of multi-channel EXRs
-No Mac version
-Fewer job opportunities (at least in the vfx industry)

Both are very robust and professional composting programs. Although Nuke is the defacto standard in vfx production these days, I know for a fact Fusion was the primary composting platform for MANY feature films and television shows. Eyeon kind of dropped the ball for a while and let the Foundry take the lead in this regard, but I'm so glad BMD has decided to breathe new life into Fusion and give Nuke some much needed competition.

-Mitch

Here's another thing I find puzzling about Fusion. With the exception of Unchartered Territory, I see very few "pro-level" samples of Fusion work.

Having used both in production here is my quick comparison list.

Fusion Pros (vs Nuke)
-Overall faster rendering
-More intuitive
-Better 3d toolset
-Superior text/titling tools
-PRICE!!!

Cons
-Clunky management of linear color workflow
-Poor support of multi-channel EXRs
-No Mac version
-Fewer job opportunities (at least in the vfx industry)
...

-Mitch

I should add in the Cons :
- No REAL deep compositing ( the use of DCM )
- U.I : You click more in Fusion than in Nuke

and in the Pros :
-Channel boolean

I've used both Fusion and right now I'm using Nuke, and I can say that I'm very happy for the fact that Nuke has a very good competition right now with Fusion under BMD. Although The Foundry is far from being lazy developer, sometimes I wish they would do few things just like in Fusion. And lower their prices, for realz

Eyeon didn't give a sh... about marketing and PR, thats why you cant find pro level samples reel.
That was one of many reasons why fusion lost position in pro Vfx work
Hope BM will make fusion reel soon.
There was nice topic on pigsfly forum about Nuke vs Fusion, with conclusion that you can do the almost the same in two programs. I have tried nuke demo and i didnt like it, way to technical. Fusion is more artist friendly and easy to learn.
I dont like the idea of fusion as replacement for motiondesign after effect. Fusion is compositor just like nuke.
k

I think this is mostly the fault of poor marketing by Eyeon. Frantic Films, Digital Dimension, Blur, Eden, Flash Film Works, Rainmaker Vancouver and may other studios used Fusion as their primary comp package for years. Here are some examples of work done using it...

My personal reel (a bit outdated). 95% of what's in here was comped in Fusion


Digital Dimension Reel. Most everything here except "The Grey"


A friend's reel who worked at Frantic / Prime Focus. Majority Fusion shots.

Each tool has its place. Since I own both programs I am very anxious to see where BMD takes Fusion Studio. The biggest thing Fusion is missing for me compared Nuke is native Linux support which is a huge deal for me since my main workstation and render farm both using CentOS Linux.

Please tell me how Nuke is resolution independent, for example.

I read somewhere Fusion does up to 32k x 32k, no?

I just rendered a 32001x32001 Pixel image, was no big deal. When I set the project settings to much higher, like 150k by 150k, it clips the viewer to 65536x65536. This might be a limitation to some, but I think till you need that resolution, well...

Hi,
i'm sorry - i think You're missing the point, as most people here.
there are lots of Fu vs Nu threads in other fu-related forum; and please don't go THAT way.
there are just too many of differences. some of them can be implemented / workaround-ed; some not

Resolution independent-wise:
Nuke, like AfterFX, stores position data as pixel value (ie in 1920x1080 picture, the center position is 960x540). so when You switch to different format, the position stays the same. This is even the issue when You switch-on Proxy-in-the-Viewer and draw the mask (nuke 8v5). after switching the proxy off, the mask doesn't resize itself info full rez.

Fusion stores position data as a percent of a picture / resolution, You're working with. So if You scale the footage, every element will stay as it was.. But it also has some drawbacks - mainly from implementation method in paint/tracker, imo.

cheers

Thanks for unlocking this one!!!

@Mitch, thanks for those links -- very nice Fusion samples. I've also been watching the Rony Soussan YouTube videos. Very impressive. Really looking forward to getting into Fusion now.

Thanks for unlocking the thread.

I feel that this thread does very much belong in this forum. Let me explain...

Many of us are looking on curiously how Fusion develops now that Blackmagic has taken over. One of the main concerns for us Aftereffects users is the learning curve which seems steep no matter how one looks at it.

I think I speak for many of us when I say that we'd all prefer not to have to learn both Nuke and Fusion...so it helps to have a discussion and explain the strengths and weaknesses of either app. While it's unlikely that either is a clear winner, it is useful to know which is better at what. Of course, the price structure of Fusion makes it difficult to ignore regardless of its weaknesses....however right now the lack of a Mac OS version is a challenge for sure.

Anyway, thanks for reopening the thread.

Open CL is Rocks! now I have 2 Amazing software for post production Fusion & Resolve
my GC (AMD Pro Fire w7000) is alive no more After effects thanks BM

Get both programs if you can afford to and Resolve while your at it for good measure. You really can't go wrong using Fusion Studio, Nuke(whatever version), and Resolve all in the same pipeline. I know I will be!

Resolution independent-wise:
Nuke, like AfterFX, stores position data as pixel value (ie in 1920x1080 picture, the center position is 960x540). so when You switch to different format, the position stays the same. This is even the issue when You switch-on Proxy-in-the-Viewer and draw the mask (nuke 8v5). after switching the proxy off, the mask doesn't resize itself info full rez.

Fusion stores position data as a percent of a picture / resolution, You're working with. So if You scale the footage, every element will stay as it was.. But it also has some drawbacks - mainly from implementation method in paint/tracker, imo.

I see, I think "Resolution Independent" is easily to be misunderstood in the context, haha

Though I have to say, I don't really see the problem for that to be honest. In AE, sure, in Nuke - not so much.
The only time I came around such a problem was when working on 2K source clips, and then getting the graded Clips later in HD and I have to apply the changes accordingly. So any Painting and Tracking etc does not fit anymore, naturally;
This can of course be fixed when you know up front the Grade will only be HD, or if you can ask the Colourist to just render at Source-Resolution. My workaround is a simple Reformat node though

Never had another case where this was a big problem, but I see that it could annoy at some point

I'm a Nuke user looking to give Fusion a try. Are there any guides or tips for Nuke users transitioning to Fusion?

When Shake was discontinued and a lot of Shake users moved to Nuke, there were things like a Node conversion guide that helped Shake artists understand where their "go-to" nodes were in Nuke. These were created by other artists but I'm just wondering if anything like that exists for the Nuke to Fusion transition.

i dont think this exist (yet), because there was only a change from fusion to nuke before.

but if you have specific questions feel free to ask.

biggest difference will be how you work with multichannels, this is not really there in fusion. if your used to create and reuse dozen of channels in nuke you have now to rethink alot and work more like you shuffled everything

Please tell me how Nuke is resolution independent, for example.

I read somewhere Fusion does up to 32k x 32k, no?

I just rendered a 32001x32001 Pixel image, was no big deal. When I set the project settings to much higher, like 150k by 150k, it clips the viewer to 65536x65536. This might be a limitation to some, but I think till you need that resolution, well...

You can render all sorts of crazy sizes, yes. But not all of the tools in Fusion support resolutions beyond 8K, while others have issues beyond 32K. But whatever you can get away with is fine, just know that there are limits. Same thing in other applications too.

BUT... you're missing Pieter's point. In Fusion, nearly all coordinates in the UI are expressed as a percentage of the image width. So a blur, for example has a radius that gets bigger or smaller based on the size of your pixels relative to the image. So it's the "same" blur when you are doing SD or 4K. This is what lets the proxy system work so well. You can build a comp with HD JPG proxy footage, then swap it out with 4K originals and everything just works. In Nuke, you have to rebuild your comp. So in that way we say that Fusion is resolution independent because it doesn't actually care what resolution the input is, the tools all scale themselves appropriately.

- Chad

Indeed, the situation with examples of big film projects composited in Fusion could be better.
And it's not because the software was rarely used, but rather because it always takes quite an effort to get the rights for publishing the imagery.

Therefore here are my 5 cent:
vimeo.com/72041693
This reel is made of shots I completed as the Lead Compositor at Parasol Island GmbH
All the compositing, color grading, all particle effects apart from drops on the car, all volumetric effects apart from the clouds in REWE and Svedka were done solely in Fusion.

Also, in all the film projects listed here:
imdb.com/name/nm3054479/
Fusion was used for compositing.
Fusion's amazing particle tools were used in all of them apart from the 9-th Company.

what kind of application use left click and middle click for zoom ? this is the reason i stay away from fusion,.. gone are the old days of 3 mouse button. its left button scroll button and right button generation.

Who use mouse these days ? Go get a tablet, an enjoy Fu

Yeah the mouse needs to be configurable like any 3D app, keeping a treamlined workflow between packages in this way helps workflow so much.
The whole "get a tablet" thing is getting a bit tired, we like the mouse, problem with that ?

"The whole "get a tablet" thing is getting a bit tired, we like the mouse, problem with that ?" No I don't have problem with mouse user at all. But we are all tired about this complain about mouse nav too..
This thread isn't about navigation in 3d space, but Fusion vs Nuke. Add your wish to the wish list.
peace

This thread isn't about navigation in 3d space, but Fusion vs Nuke. Add your wish to the wish list.
peace

Ironically you're the one who recommended using the tablet. Also, Fusion vs Nuke pretty much insinuates a wish list of features. It's a given.

I'm hoping for a robust camera tracker in Fusion like Nuke (a vs and a wish). I may catch junk for this but I do hope the UI gets streamlined and flattened, to keep up with modern UI/UX.

I may catch junk for this but I do hope the UI gets streamlined and flattened, to keep up with modern UI/UX.

I hope the UI gets user tested (yes, even A/B tested) and customizable, since modern UI/UX is basically a crapshoot. Old versions of Fusion let you adjust the skin so you could work in a way that was most comfortable and efficient for you.

A quick note about navigation - which I do think is relevant to such a thread - in Fusion don't bother with trying to use the LMB/ MMB combo for zoom, just use the ctrl scroll wheel instead. More intuitive in my opinion.

Of course if you use a tablet pen, the LMB/MMB would then work well.



Nice Reel Mitch!

Hi eveyone,
I know this is an old topic, but it's perfect for my question, that isn't actually "Fusion vs Nuke which is better", but "Fusion vs Nuke which is better if your main software is Davinci Resolve?"
In other words: the bridge Davinci/Fusion gives something more than the couple Davinci/Nuke?

Thanks for your advices.

I wouldn't shift the argument of FvN on Resolve. Artists have been working just fine w/Nuke and whatever formats are given to them. If anything, with NS, there may be more flexibility. If there is a benefit to some sort of workflow between Fusion and Resolve, it's negated by the feature set comparison between Fusion and Nuke.


I am curious as to if BMD will show Fusion 8.5 or 9.0 at NAB this year. It's been relatively quiet on Fusion releases over the months.

If Fusion Connect gives you enough leverage to choose Fusion over Nuke, I would say do it. And I would also guess you are not a compositor but rather a colorist/editor/etc who has to do some occasional comp related stuff. For compositing, the "worlds best compositor" has to get even better, at the moment it simply makes (my) head hurt.

So what is it you're missing, Hendrik? Maybe we can help you out.

So what is it you're missing, Hendrik? Maybe we can help you out.

The devil is in the details as it is said. It all starts with the UI and interaction with the tools that you do a lot and uncomfortableness piles up. A very simple example: the float knob.

Float knob is presented to the user as a slider and a value. Seems nice and intuituve allright. Now, what is the range of the slider? You can't tell because it doesn't have associated value labels on it. So slider is only useful for sliding and peeking with your eye, what the value box says. You can't click on the slider to get a fast approximation for what you are after because a) you don't have a clue where to click and b) slider jumps in steps, not to where you click, wtf?

The value box of float knob is no better. It displays the value, but to change it, your only way is to punch in new numbers. There is no way of incrementing/scrolling the existing value (think of minor adjustments for matching color values in different elements etc). And on top, after punching something in, active knob changes (jumps to next editable value) so to make a new change, I must manually make it active again. This is where my head starts to hurt.

The advance to the next setting /knob/value is a preference thing. You can change this

Float knob is presented to the user as a slider and a value. Seems nice and intuituve allright. Now, what is the range of the slider? You can't tell because it doesn't have associated value labels on it. So slider is only useful for sliding and peeking with your eye, what the value box says. You can't click on the slider to get a fast approximation for what you are after because a) you don't have a clue where to click and b) slider jumps in steps, not to where you click, wtf?

The range of the sliders is dynamic. And on many sliders it's not linear. So labels wouldn't be useful and would increase clutter.

Clicking on the slider is equivalent to a spinner. That way you can click it 3 times and each time it moves the same amount. If the slider moved to where you click, you'd only be able to click it once to get a useful result.

There is no way of incrementing/scrolling the existing value

See above.

It is to some level a matter of preference, but I speak from practice here, not theory. Somehow at least one nother software has managed to accomplish these things without clutter and major pita and I can't find any logical benefit in the way these things are handled in Fusion. But, as I said, it is a matter of preference and pita tolerance.

It is to some level a matter of preference, but I speak from practice here, not theory. Somehow at least one nother software has managed to accomplish these things without clutter and major pita and I can't find any logical benefit in the way these things are handled in Fusion. But, as I said, it is a matter of preference and pita tolerance.

I just explained the logical benefits. You asked how you can increment a value and it's built into the control by just clicking on it. It's not a preference issue, it works better to make the clicks relative vs absolute.

So, if I want to get a fast approximation of the value I am after, then want to change it in 0.1 and 0.01 increments and after that really drive it home in 0.001 unit increments, how can I accomplish that without digging in some preference setting every time my level of precision needs to change? This scenario is very common when fitting different elements together, for example clone patches. I do a gamma slam on viewport and then need to precisely match the patch to bg in every color channel.

Page 227.

The other pages are useful as well.

Coming back to the example, using a ColorGain op I want to adjust the red gain. Clicking increments in steps of 0.5, shift-clicking also 0.5 and ctrl-clicking in 0.008 if going past default value. When on smaller side of default, steps are 0.1, 0.1 and 0.002. I have a hard time finding logic behind this. Is it possible to change the increment step preference?

Because it's a non-linear slider. One stop is *2 or *.5. So it makes sense that going down you'd have smaller increments vs going up.

Yes, you can change the setting, just use "Edit Controls..." to set the center of the slider to 0 instead of 1 and save that as the default.

No you can't set it globally in the same way you can't set all the sliders in the application to be integers or to have fixed steps or have a minimum of zero. Each control has default settings that are there for a reason and global overrides don't make sense.