[leif3d]

Will there ever be a proper Linux release? QT maybe? or will we be always constrained to a wine reliant 32-bit version?
Our shots right now are very heavy, therefore 32-bit architecture is not an option for us.
Release after release I hope for a native Linux build in 64-bit, but it never comes.
Any road-map would be nice.

[leif3d]

I guess noone cares about a native Linux version... :i
Very strange now with Nuke being so popular and Maya Composite gaining some ground. Both have a native Linux version.
I've never been able to get a straight answer from eyeon as to why they don't release it... specially with so many houses being Linux (Maya/PRMan) reliant.

[pingking]

i think this was discussed before and the answer was:
it is a native linux version, the wine is only needed for some plugin compability

but i never worked or tested the linux version, so i cant say anything about it

for a QT version eyeon would need to rewrite the whole application and a second total rewrite within 5 years sounds not very likely, but i would love to see a more flexible GUI

[leif3d]

Thanks for the response Pingking,
By native I mean non WINE reliant, and by non WINE reliant I mean I want a 64-bit version of Fusion on Linux.
If it were completely native it wouldn't need WINE at all no? How about if I don't want Quicktime and AE plugins? Right now there isn't a 64-bit installer...

[Tilt]

leif3d, on 08 July 2010 - 12:56 PM, said:

Thanks for the response Pingking,
By native I mean non WINE reliant, and by non WINE reliant I mean I want a 64-bit version of Fusion on Linux.
If it were completely native it wouldn't need WINE at all no? How about if I don't want Quicktime and AE plugins? Right now there isn't a 64-bit installer...

Wine is for the GUI (might also be needed for the plugins, as pinking said). The rendering code is native.


http://vfxpedia.com/...?title=Why_Wine

The 64bit thing is a different issue, maybe ask eyeon support or sales directly.

[hulan]

Tilt, on 08 July 2010 - 03:42 PM, said:

leif3d, on 08 July 2010 - 12:56 PM, said:

Thanks for the response Pingking,
By native I mean non WINE reliant, and by non WINE reliant I mean I want a 64-bit version of Fusion on Linux.
If it were completely native it wouldn't need WINE at all no? How about if I don't want Quicktime and AE plugins? Right now there isn't a 64-bit installer...

Wine is for the GUI (might also be needed for the plugins, as pinking said). The rendering code is native.


http://vfxpedia.com/...?title=Why_Wine


The 64bit thing is a different issue, maybe ask eyeon support or sales directly.

I was checking on the wiki for Wine 64bit (http://wiki.winehq.org/Wine64).

Quote

This port is just beginning to work, but is not yet fully functional.

Would seem that we shouldn't hold our breath for a 64bit version?
Eyeon should consider doing the same thing as Nuke did from Nuke 4.x (I belive) to 5 and do the UI in a cross platform framework such as Qt4, which is proven stable on Windows, Linux and OSX.

[SirEdric]

Hi Guys.

Sorry to disturb your dreams a bit...
But Qt is tremendeously slooooooooooooooow.

Do you like the way Fusion interacts in real time to UI commands?
(i.e. you drag any control and get instant feedback in the view)?
I bet you do, since this is something that always has been one of the unbeaten strong points in Fusion over the last 20+ years.

Tests have been made with Qt, and it's far inferior to the way we handle UI/View/Rendering currently.

So...yes...Multiplatform would be nice (and will come eventually....;-)) but Qt is not in the closer choice for now.

Cheers.

Eric.

[bfloch]

I can't really second that but it all depends on what and how you test.
I recently wrote a ImageSequence viewer in PyQt (!) and it instantly draws frames AND overlays if you move the slider. Works well for 25 fps of cached HD images on a workstation and SD on a Laptop. OpenGL should be not related to it at all and I doubt that redrawing a slider gives much of a problem. Qt is no FLTK, but that's why I like it. It makes even C++ coding fun.

But there can be definitely speed problems. Especially when compared to native drawing. But this doesn't help the portability. It is not a coincidence that Nuke's Node Graph drawing is OpenGL. The foundry also said in the interview/feature on the qt homepage, that Qt is big and has a lot of stuff they don't need. It is a pity that the QtGui and QtCore dll are about 15 MB if you want to distribute a little python code as executable.
But if you take the benefits of it and take the stability you won't find much alternatives. Has been proved to work in our industry. Else companies like The Foundry, Maxwell and most recently Autodesk made a HUGE mistake. I suppose they also test different frameworks. And there must be reasons for Qt, even if speed is not one of it.

wxWidgets is definitely a nice concept but from what I saw you end up having to write much platform dependent code, and that is not very efficient.
Though I still leave the decision to you pros. Just don't make your lives harder that it is.

Yes customers reported that the UI is slower when Nuke hit version 5. But no no one cares today and I bet the Foundry coders have fun working with the Qt framework.

My 5 eurocent.

[bfloch]

BTW this is similar like Lua vs. Python. From speed point and threading etc. Lua is probably the winner. But that's not the selling point.
Cheers to python. And cheers to whatever you may use for Linux. I would love to use the current Fu Linux, am not afraid of wine for the UI, but since I found out that my dongle is not sufficient :-/

[leif3d]

Thanks for the straightforward response guys.
I'm not experiencing any slowness with Maya's QT interface under Linux when compared to Motif, but it IS pretty slow on windows and OSX. I'm not sure if this is a Qt or platform issue though.
Regardless of the slight slowness that we might see in the UI (according to Eric), the benefits of platform independence and stability will far exceed this minor inconvenience. If QT also gets rid of disappearing splines and glitchy timeline, add another vote!
A similar choice has been made with Python, which makes me very happy, because I'm learning Python for Maya. My productivity just went up because of this great step. I can only wish for a slick Linux version of Fusion.

One of the reasons I still have a windows box is for Fusion. Please help me get rid of it... :)

[bfloch]

leif3d, on 08 July 2010 - 05:30 PM, said:

One of the reasons I still have a windows box is for Fusion. Please help me get rid of it... :)

Me too. And I am really excited about EditShare LightWorks. Finally a solid editing solution for Linux? And free as freedom beer? Or just a marketing gag?

Btw. wouldn't it be easy to port Generations? Isn't it 90% OpenGL?

[ChadCapeland]

leif3d, on 08 July 2010 - 05:30 PM, said:

Regardless of the slight slowness that we might see in the UI (according to Eric), the benefits of platform independence and stability will far exceed this minor inconvenience.

Really? Inconveniencing 100% of your customer base to allow a minority to avoid running WINE? That sounds like a terrible tradeoff.

- Chad

[bfloch]

We're talking about probably number one feature request after DoD, not to mention that this is a totally new market for professionals on the Linux platform and a lot of unsatisfied After Effects dudes on Mac osx. In that regard the current Fusion customer base is the minority.

This sounds more like being afraid of changes. Most companies run different platforms. Visual Effects, unfortunately is defined on multiple platforms. I bet that for a lot of companies that switched from shake to nuke this was one of the main reasons not even to try fusion. Offering a Linux Admin a Wine app is like giving a vegetarian meat to eat. And it is not even stable from what I can tell.

And it is not said that porting to a multiplatform framework will make fusion worse at all.

Actually from my personal view there are two main reasons why Nuke is the top app on big sites. The Os for technical reasons and the scripting and customization API. You only need to watch their python videos to realize how easy and perfectly it can be integrated in a pipeline. And fusion is not that bad on this. Even got 200% better the day we got the PeyeonScript Lib. But there are certain things that I miss (next to the docs :P) like custom UIs and callbacks that work from within the app and tools/nodes. Seriously it is not that bad with Fusion and UserControls, simply if you study it Nuke made most of this smart customization possible in 4 to 5 and 6. They didn't pack that much features, and seriously I believe that feature-toolset wise fusion is still #1. But this two factors made it not an option for most companies in the industry.

So I can't share your opinion Chad. We are not talking about a minority. This is probably the biggest hurdle to take on a software. But unfortunately it is necessary in VFX. I can feel you though Chad. I too would be happier if we all used Linux instead :P

We can run a poll on this if you feel better but I don't think this is necessary since fusion management or marketing should know.

[ChadCapeland]

bfloch, on 09 July 2010 - 01:52 AM, said:

We're talking about probably number one feature request after DoD, not to mention that this is a totally new market for professionals on the Linux platform and a lot of unsatisfied After Effects dudes on Mac osx. In that regard the current Fusion customer base is the minority.

Of course non-users will always outnumber users. That's why Lotus is busy designing a new box truck, because the market there is so much larger. But you could argue that Eyeon should make a special version of Fusion targeting law enforcement, or life sciences, or data viz. Larger markets there, too. But there's only a finite amount of work that can be done in any given time period. Having all core, SDK, and UI development stop for 2 years would kill Fusion more than releasing a new UI in 2 years would save it. Hiring a new "Qt" team to do the migration while development continued for 2 years would add $500 to the cost of subscriptions, and that's going to cost you users over time, too.

bfloch, on 09 July 2010 - 01:52 AM, said:

This sounds more like being afraid of changes.

No, it's fear of uncertainty. There's no assurance that a re-write of this magnitude would bring ANY new customers, and could bankrupt Eyeon, or at the very least stall them from improving the product to the levels we need.

They're a business, we're a business. If they screw this up, both businesses suffer (though Eyeon would suffer more, we've got a lot of time/money/code invested in Fusion, too).

WINE provides a low risk way of running a Linux renderfarm. I'm sure if there were a lot of customers who were actively working on Fusion in a workstation environment, things would improve. But there isn't pushback from users, not that I've seen. Just sideline/backseat stuff. It's far easier to take the simple approach and spend $200 or less getting Windows on your workstation and knowing everything will just work. And yes, for VFX users, Windows is still the dominant platform.

bfloch, on 09 July 2010 - 01:52 AM, said:

And it is not said that porting to a multiplatform framework will make fusion worse at all.

No one said it would. Just that it likely would. It's risky. Would developing a 3D tracker be risky? Optical flow framework? A solver (ala Scene Genie)? Image synthesis? Better connectivity to 3D apps? Editing systems? Updating the SDK to support newer (and fewer) VS? Documentation? The wishlist of serious projects that are more likely to be successful is already quite large. The wishlist for individual tools is already huge.

bfloch, on 09 July 2010 - 01:52 AM, said:

We can run a poll on this if you feel better but I don't think this is necessary since fusion management or marketing should know.

Poll might be useful if Fusion was open source. :)
As it is now, only a couple of people have opinions that matter.

- Chad

[bfloch]

I can see your point Chad.
I think this is like in every business. Someone has to take responsibilities for actions. For the benefit of the product.

If I were in charge my thought would be:
Fusion claims to be a Film compositor. Look what the Film dudes use. If Fusion targeted more towards Motion Graphics it could be different. Or at least OSX not Linux.

I do follow eyeons decisions but am not always happy with them. I hope they do the right thing for their customers or the majority of them or at least for future customers and their product. But too me there are signs that they are not always doing the right thing. It's not always aggressive marketing when you loose big customers.

The big question that no one can answer but eyeon is what way to go for the future. And once they have a Roadmap it would be nice to share it with their customers.