Fabric 1.12: GPU Compute, Splice Rendering, Characters

[Pooby]

Yes the generic painting toolset looks extremely exciting.

I'm also particularly keen to see the visual programming that is being introduced later in the year.

From the Fabric Site

Later this year we will also be offering visual programming capabilities for Fabric Engine, so if you’re most comfortable in environments like Softimage ICE, watch this space!

[Sil3]

This means you can build your tools faster, fix bugs without waiting for AD, and add new tool types as well.

And for people like me that "dont build" when can we expect to see something we can play with? Or its just a TDs wet dream for a long time to come?

[McNistor]

Not my intention to rain on anyone's parade, but I see this as a purely TDs tool and not an artist solution.
But just as one of the Maya's touted strengths is "openness" this is just as problematic for someone without at least a few coders/TDs in case of Fabric.
Hell, if "openness" is something to be put on a pedestal, then we ought to put there the notepad as well since I can open it anytime and start writing code and saving as a .cpp or whatever.
I'm not saying Fabric was ever advertised to be something else so I actually don't have anything against it or other things like it, I just wanted to say that for the artist/generalist, pre-built solutions are a must.
That's another way of saying looking forward to the day Fabric will become an actual app. rather than a platform/env/whatever it's called.

[Manticor]

Im kind of hoping for the community to grow and a collaborative effort between people who can use fabric to build a new dcc between them .A good base framework for things to hook into with a basic modelling system to start with would be ideal .But it still too early in my opinion as a lot of things are still in flux or are just missing at this point...which is making me hold off from really getting into and learning it properly .But having said that I am highly optimistic about its future ,the dev team are legends and the mailing list is like a who's who of the xsi td elite.
I also think its about time they had a forum as well as the mailing list.

[EricTRocks]

One thing I think people need to be clear about is that Fabric is a platform, API, framework to build tools with. They are implementing features that will allow TD's / Developers to create tools needed at their studio / company. Keep in mind that Fabric Engine isn't just for VFX. It can be used for medical imaging and other computationally heavy needs. They all come from a VFX background but are building a framework that can reach VFX but a lot of other industries as well.

So currently, yes you need TD's / Developers. But is that a bad thing? TD's and Developers need jobs too. Especially those who are Softimage inclined. It's hard for us to find jobs now. It's the same thing that ICE had going for it. It was super useful and handy for the really advanced users to prototype in (then implement in custom ICE nodes in C++ or something else), helpful for the moderately skilled TD's to build setups and tools using it to get production work done and solve problems that couldn't otherwise, and for the technically minded artists, it offered an easy jumping off point to learn and get more technical.

For the average generalist and artist who just push buttons in the interface and who don't dive in to the technical world, you're probably not going to use it on your own. You'll use the tools made by a TD.

Right now I could write a tool that clones Chronosculpt in Fabric Engine. It would be able to run in Maya, Softimage, and stand alone. That could happen right now if someone chose to do it. Someone just has to do it. There are much simpler tools that could be made as well. Since you can save out the *.splice files there is nothing stopping someone from passing around an IK solver that simply takes 3 control inputs and drives 16 deformers for the arm. Provide a *.splice file and a script and done.

The community needs to grow and it is, but people also need to be willing to either get off their ass and learn some advanced skills and not just be a button pusher, OR just be patient / pay someone to build the tools you want.

That being said, I'll leave it to the Fabric folks to comment on their goals for 2.0 and whether they are going to provide an "artist friendly" environment / tool set.

[Pooby]

Fabric is absolutely for TDs. Don't expect fabric to directly make artists tools. It's not part of their claims.

However, TDs are there for artists. And we will be the ones using Fabric to make these tools.

[Hirazi Blue]

I think most people fail to recognize that ICE is nothing but a framework also, visual in nature and as such easily accessible, but in a way a "tool for TDs" just the same.
;)

[Bullit]

Well ICE is a Frankenstein. You can see - even i official v7 instructors course - mostly ICE stuff done for artists not TD's.

I sometimes think that a clearer split like what was achieved with Max Particle Flow was not better, at least i think the nodes and functions would be more organized even if the interface could suffer.

[EricTRocks]

Well ICE is a Frankenstein. You can see - even i official v7 instructors course - mostly ICE stuff done for artists not TD's.

Well, I'd say it really didn't solidify for artist friendly workflows until they implemented the ICE module in the UI and the 'Simulation Root' node. Look how long it took them to do that and work that all out. They had to make sure the ICE system was working in a decent fashion and had to set aside dev time to build the artist friendly tools. Same needs to be done with Fabric.

Also, look how other devs like Eric Mootz are doing great tools on top of the ICE framework. The Softimage 3rd Party Dev world was / is tiny in comparison to say Max and Maya. Seems that instead of paying some 3rd Party Dev to build some tools, people would rather just wait it out until if and when it gets put into the Software.

3rd Party Developers for Softimage had a place in our community but were quite neglected and in my opinion need to be supported a lot more in the community that we evolve into.

[Bullit]

I think it isn't and will for obvious reasons never be solidified for artists despite most official tutorials being for artists. In short you can do much less with what you call the artist interface of Softimage than with Particleflow where there is a more clear cut between Artist and TD part at expense of interface.

This dichotomy is another zig zags that Softimage management did. ICE is for TD's, but promoted for Artists. One idea could be to try promote artists to TD level but that was never done.

[Pooby]

I think it isn't and will for obvious reasons never be solidified for artists despite most official tutorials being for artists. In short you can do much less with what you call the artist interface of Softimage than with Particleflow where there is a more clear cut between Artist and TD part at expense of interface.

This dichotomy is another zig zags that Softimage management did. ICE is for TD's, but promoted for Artists. One idea could be to try promote artists to TD level but that was never done.

That's exactly what ice achieved. I know many people who couldn't code, didn't know what vectors were etc before ice came along. Who now are making tools.
I dont like this artist/Td distinction. Many of us are both and anyone could be.

[FXDude]

I remember my first ice tutorial, (Bradley Gabe's flaming hand video)
and before knowing anything about ICE, I could understand everything that was happening (watching in awe).

And recently when checking-out Houdini, I stumbled upon an Approach to making the dandelion from Marks Shoenegel's demo using LaFrourmi's "Car Parts" production assets.
(both of similar end-result complexity as Bradley Gabe's tutorial)

And it was -absolutely not- the same experience at-all (not because of the poor audio, or my lack of houdini knowledge)
just looking at all the steps (and the kind of steps) involved.

It was painful to watch even while now knowing the principles
especially now being use to something made to make complex things simple.

Making complication Simple is not just for when actually learning,
it is also for when actually doing (and dealing with) things (fast), deciphering things made by others, or deciphering yourself the next day,
and is proportionately the single most difficult /time-consuming thing for developers,
as well as the single most time -saving- thing for -relatively- technically literate experienced end users.

But I hardly think that's what kept FE from making the jump to becoming more "artist friendly",
FE has been hovering just below the line of actually becoming a full blown DCC for a while now,
and if that last step was not made, I'd be surprised if that was by actual limitation.

I imagine that may have to do with the "dictates of the market", and I know going against that may not be easy or safe.
("market dictates" also being the cause of one of the most workable platforms making way for one of the most awkward standards ever,
thus creating a pretty incredibly stark contrast, making any somewhat more elaborate setups comparatively more reserved or less accessible by a .. pretty incredibly stark contrast)

For me, I personally don't think the bulk of doing High-end stuff (along with involved problem solving) should (necessarily) be a task -mostly- all about minding syntax or debugging code.

Which is exactly the main idea behind SI's design at it's very core (ICE or otherwise) .. being all about "making accessible" (or at-all possible) what is normally.. not so much (to this day) , allowing to do elaborate custom things VERY FAST, expanding the reach of both artists and the most technical artists alike.


Yep.. the word "regression" comes to mind.

[Bullit]

I agree that it achieved it in some people. You are a good example of it. But it wasn't a big impact. A proper way to do it was Avid at time recommend books and other learning material if it was unable to do it itself.

I think the distinction is relevant to show those that know the guts of 3D applications and those that don't, but calling someone a Technical Director seems to imply that person doesn't have any artistic capability so it is a description that might be not wise.

[FXDude]

I think the distinction is relevant to show those that know the guts of 3D applications and those that don't, but calling someone a Technical Director seems to imply that person doesn't have any artistic capability so it is a description that might be not wise.

Yes a "TD" label can indeed imply that, and I think it's pretty clear (if not encouraging) to see that people working on FE are very-much both artists and TD's :]

[Hirazi Blue]

I think there is a missing link between "the TD" and "the Artist" and that's specifically a group we Softimage users have gotten a lot of - mainly thanks to ICE, I guess - and that's the also vague category of "the Technical Artist", someone who isn't primarily technical in nature, but is willing to get his/her hands dirty ("icy"?) when the need arrives.
;)

[Draise]

I tend to agree with all yous...

I think a good way to market Fabric Engine is to be something of a TD world with Artist Friendly implementation. I like how open and powerful the potential of the software is and Splicing is, but sometimes I just want to use existing scripts and code that can be modified to what I need the most - applied technical work of others from a community to an artistic end.

If you had FE with already artist friendly, useable systems plus the ability for the artist to get their hands dirty by being more technical - with the added option of full access to coded development - it could be just the perfect artists/technical generalists tool to take a step up from in-the-box packages.

Like Softimage's Feature Set in ICE: with it's ICE rigging, Crowds, ICE syflex, particles, and then Eric Mootz plugins, scatter tools and other compounds by third parties: FE should come with a set of user artist friendly and TD ready-to-develop applications and artistic examples - tools.

Examples could be:

• GLSL shader library or compounds with realtime SSS, Blinn, Phong, Fresnel, transperency, etc?

• Realtime viewport Post Effects (DOF, Bloom, GI, AO - etc)

• Cloth System?

• Strand/Curve system?

• Particle System?

• Fluids Systems?

• Rigging architecture?

• Animation enhancement toolset?

• Sculpting/Painting toolset?

• Geometry Deformers?

• Procedural Geometry?

• On the fly Geometry generation?

• Voxel Geometry?

• Crowd system?

• Node GPU enhanced compositing?

• User friendly UI/overlay and Synoptic Creator?

With simple showcase artistic "applications" built in and shipped with it, the FE could open doors to inspire newer and improved community plugins that could be bundled in with the package. TD's could modify, expand or create their own systems like they would with ICE or the current standing system, but this time around it would be motivated and pushed by the artists who would invest in a system that enhances their tool of choice - for all the systems it could technically Splice into.

I think that is what they want really in the end... but it's all still very "technical". I myself know very little of scripted logic, and use ICE in a limited fashion, but thanks to ICE I am starting to get into that concept of developing "logic" a bit more to get where I want to go artistically - be it in SI or in the other packages I use. I can't wait till they have nodal implementations, or more: template compounds/tools I could fiddle with from the get go - as I am more of a lego person than a "create from nothing" scripting person. I think FE is on a good track if they promote a "lego" like additional feature blocks system for any tool it splices into.

I think it's worth it if it empowers the Artist, facilitates the TD, and unifies the separation between a developer and an artistic implementer, software to other software - maybe when it's nodal system and more tutorials/showcases on artistic implementation, pipelines and tools (and even user friendly tools on the surface) come out, it will be even more worth it's bang for it's buck.

And I'd then invest in it for sure when I can - me and my small team of merry men.

EDIT: it has a lot of these features already, actually.