current Fabric Engine usage?

Forum for users who have migrated or are migrating to the Fabric Engine
EricTRocks
Moderator
Posts: 754
Joined: 25 Nov 2009, 01:41
Contact:

Re: current Fabric Engine usage?

Post by EricTRocks » 09 May 2015, 17:33

Not likely we'll share the tools. It's what gives the company an advantage.

Kraken will be open source and will come with basic solvers that will allow anyone to build a biped. However the Hybride solvers stay at Hybride.
Eric Thivierge
Lead Kraken Developer, Fabric Engine
http://fabric-engine.github.io/Kraken

User avatar
Draise
Posts: 891
Joined: 09 Oct 2012, 20:48
Skype: ondraise
Location: Colombia

Re: current Fabric Engine usage?

Post by Draise » 09 May 2015, 18:03

Propriety stuff ... just trying to say... it would be nice to have at least a testimonial documented on their site and a library of "knowledge" or articles or "what can be done" reviews.. saying it CAN be done, doesn't have to be shared persay... just thought "opensource" without having to share the source. :-bd Sorry for sounding greedy.

-Marketing FE so that more people get into it, more talent, man-power, people getting encouraged to work with it, and thus.. probably making it's sure of it's survival in the industry (unlike SI).

Thanks again for sharing what you do with it. :)

EricTRocks
Moderator
Posts: 754
Joined: 25 Nov 2009, 01:41
Contact:

Re: current Fabric Engine usage?

Post by EricTRocks » 09 May 2015, 18:20

Believe me, there will be quite a bit of info coming out about the work we did. We presented at the Fabric User Group at Siggraph last year and will probably be there again this year showing stuff we've done.
Eric Thivierge
Lead Kraken Developer, Fabric Engine
http://fabric-engine.github.io/Kraken

User avatar
MauricioPC
Moderator
Posts: 1085
Joined: 16 Sep 2013, 13:39

Re: current Fabric Engine usage?

Post by MauricioPC » 09 May 2015, 19:24

Thanks for all the info Eric. The good thing about this is that if Kraken works on all software that have FE, we'll probably see more cool stuff done on animation outside of the Autodesk helm and that can't be a bad thing.

Modo animation still have problems because of lack of tools and viewport performance. I imagine that that crazy fast viewport from FE will eventually come to Modo, making it easier to use for animation. Maybe smaller studio will risk to develop something and believe in something like Modo or even another possibility. Who knows ... exciting times ahead. :)

User avatar
FXDude
Posts: 1129
Joined: 19 Jun 2012, 21:59

Re: current Fabric Engine usage?

Post by FXDude » 10 May 2015, 01:47

Mathaeus wrote:I've been witness of sooo many sad, background stories. Like producers of small team, hunting across nine European countries, for another ICE guy, once first ICE guy left the team. At the end, finishing the project without ICE. And so on and so on....
Wow that really sounds terrible, would've though it would've been easier to find Ice guys than extraordinarily talented scripters (especially then)
Numbers always were important.
I dunno. sounds to me like the "true meaning" of "prestige".
When I come to XSI with XSI 4, I believed it will spread a bit more.
Around XSI 6, it was clear for me, how entire story is going to nowhere.
also sounds really terrible,

But for the many users that were exposed to Arnorld on Maya not long after release for SI, it seemed like literally everything came from the tiny (and specific) SI community.

And before Arnold, that effervescence also went for mental ray (which was also for 3DS and also exposed to thousands upon thousands), and also goes for the more recent Redshift until today beyond EOL on RS forums. (despite all the corporate sillyness and state of things going-on)
Why?
(probably the same reason why they went first for the tiny SI community in the first place)

SI was (and is) just as welcoming or accommodating for the most technically inclined as for the "artists".


this was from the top 5 SI thread about scripting (alone)
Hidden content: [ Show ]
5 - scripting and workgroups

- Ease of scripting / making tools for even a moron like me

Scripting is so straight forward in xsi - the editor is great, the help is so accessible, feedback is great, again, 10x better than when I've done some scripting in Max/Maya/Nuke

object oriented SDK shared between Python and C++
(which makes it possible to "translate" tools prototyped in Python to C++ relatively easy)

- the API... seriously, having a logical and consistent API that is easy to use and don't need additional efforts to do simple dev tasks won't go in the way of flexibility. I don't know of any dev who didn't complain about Maya's API.

- proper and complete port of the API to Python

3. Usable Python API

1. Tool/gfxSequencer API

2) ability to overcome most problems without scripting or plugins
This is from Cristopher Crouzet (both Maya and SI User for a while)
in response to a comment a bout Maya API on his blog,
Hidden content: [ Show ]
on 23 March 2015 by SimonC
Since [Maya] API 2.00 is not complete we still need the old one, and if you're learning the API you soon discover the paucity of documentation, so it isn't easy. Learning the object model in XSI was certainly easier!!

on 25 March 2015 by Christopher Crouzet
The design of Softimage's API was so consistent and intuitive that it might be considered as offensive to compare it to Maya's! :)
About Rigging (alone)
I landed a job at Weta Digital and had to learn Maya [in ~2008]. I tried hard to approach this transition from an open-minded pespective and there was no big issue in transferring my rigging skill set. The concepts in rigging are pretty much the same everywhere. What dumbstrucked me was the process to get to that same result.

I felt like I was back to stone age. I found many of Maya’s daily tasks and the design of its API to be retarded. I spent my first weeks (months?) asking my teammates how they could possibly work with a such software. Even now, I still refuse to do in Maya a task as simple as painting some skin weights. It drives me crazy.

My colleagues wouldn’t understand my frustration. All they knew was Maya, they couldn’t compare it to anything else. I was being just annoying to them. I sometimes wished to be ignorant too—it would have helped with my zen.


But even before ice, any sort of "development" has always been like an extremely visual and intuitive experience in Softimage
and that's (in part) what seems to be missing generally.


While ICE has only merly been a (yet still a great) extention to that base.

(pre-ice video)



Personally, I learn new software all the time, and I really dont think I'm imagining things when getting the impression that working in pretty much everything to date,
is like really not the same experience at all. 3DS, Blender, Maya ... while truely awesome looking clips can be seen made using any of those packages,
which would much more be attributable to incredible user patience and determination, rather than how much software accommodates that (without involving insane amounts of time or technical wherewithal)

Mostly concerning finnickyness, figuring out, or making more than just basic things work (especially under pressure), not unlike before actually..

I think the recently posted 3DS rigging tutorial to me only confirms that, and Maya is still at the top of the finnickyness chain despite simplifications almost exclusive to the front end,
coming closer to SI like MCG comes closer to ICE, considered awesomely great for current users (getting bits and peices of what SI is all about), but faint attempts at best for SI users.


I know habit is a (if not -the- ) main factor, but I'm very much referring to beyond habit.

And if some people are inclined to stick with soft or be "emotional", then perhaps there's a reason for that, and I think it's just exactly that.


Call it being "fanboy", but it could be said that we have yet to witness what could more potentially replace it in those terms.

At the end of the day (or every day), for anyone that knows it, getting to all sorts of goals is still a (much) more fluid process.

For me, can't wait to see how BStudio is (supposedly equally friendly/inviting (with micro-accomodations and universal stable tool interop) as it is capable, but can't wait for further announcements beyond the BView ClipViewer),
cause I don't think I'm the only one to find that despite various changes making things "not as bad",
other things are still mostly pretty-much almost exactly like what they were a pretty long while ago.
(just comparatively awkward)


And for now, I'm happy that XSI can still do what it can still do best.


Concerning ICE (alone) and Canvas, maybe FE would move away from their very-much admitted "mostly for TD's" inclinations, but unless that would considerably change, even if it could technically do whatever Ice can,

(perhaps also like Houdini also technically can (and more) but just very technically)

... I wouldn't consider it to be enough grounds to say "yay! finally!"
and a similarly friendly/versatile -interactive creative environment- (not unlike the base DCC) would yet to become a reality.



(early) ICE page stuff from not 4 but 6 years ago. (page 202 of 203)

Image






"times are changing"

indeed...


Image

nodeway

Re: current Fabric Engine usage?

Post by nodeway » 10 May 2015, 13:32

MauricioPC wrote:I don't think in this case this is fair
Day spent without bashing Autodesk employee is a wasted day ;)

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: current Fabric Engine usage?

Post by Mathaeus » 10 May 2015, 18:17

FXDude wrote:
Mathaeus wrote:I've been witness of sooo many sad, background stories. Like producers of small team, hunting across nine European countries, for another ICE guy, once first ICE guy left the team. At the end, finishing the project without ICE. And so on and so on....
Wow that really sounds terrible, would've though it would've been easier to find Ice guys than extraordinarily talented scripters (especially then)
Wouldn't call it terrible, complete ICE story passed much better than I expected - today is probably irrelevant *what* I've expected. Generally I think the ratio of SI users, able to utilize ICE in everyday work, it was somewhere from one of five, to one of ten. So, importance was around this ratio.
Regarding Maya, found it nice when it comes to transforms. Skin is not competitor to SI envelope and related stuff, but like how Maya Blend shape is compact, compared to overcomplicated SI shape key system - for my taste, of course. In other word, probably second ICE compound I made, it was a ICE based morphing. Destructive, based on plain copies, but simple enough, to do not use SI shape keys, not at all, since 2008.
So, for me, main advantage of such visual programming systems is ability to 'lobotomize' or 'unlearn' the app, by replacing the overcomplicated stuff by something simpler and more predictable. Same goes to Houdini, I don't know for, and I don't care about their shelf tools - but simple 'Solver' SOP (able to turn anything into iterative, simulated process) is subject of many experiments.

Back to Maya, fortunately or not, it is *only* 3d app on planet, allowing me to keep some habits I've earned with Softimage. Just for small example, ability to key the constrained transforms back to local, even overridden SRT. Or, ability to apply the shrink wrap directly to selection of points (through 'combined' view port - out liner selection) - Blender needs some scripting to allow that. Both simply has nothing with external engines or visual programming systems, as they, by nature, have a little or nothing with interaction.

User avatar
blaxxun
Posts: 48
Joined: 04 Mar 2014, 13:23
Location: Austria
Contact:

Re: current Fabric Engine usage?

Post by blaxxun » 10 May 2015, 20:59

I wonder why nobody starts to create a very basic modeling tool with fabric.
I know... they say that their goal is not to build a standalone/monolitic DCC... also people say "Its so hard to write a modeling app".

The part that i dont get is, how can they build a complete rigging and crowd system in some months but not some basic modeling tools.
I mean something "simple" like the Softimage "N" Tool to draw polygons. For example for pbp modeling.

I still have high hopes in fabric and that they change their mind in building a complete DCC out of it.
For me i like it to have everything under one hood and i hate to have to jump from one app to the other.

Imagine the possibilities with such a modern multithreaded and gpu accelerated visual programming DCC.

Specially with Canvas which would have access to everything inside the application.
Also a sad side aspect is that tools which companys create will (of course) not be released.
I wonder how fabric will grow then...

EricTRocks
Moderator
Posts: 754
Joined: 25 Nov 2009, 01:41
Contact:

Re: current Fabric Engine usage?

Post by EricTRocks » 10 May 2015, 21:22

I'm not trying to pick your post apart in a malicious way but I think it's important to go through some of this.
blaxxun wrote: The part that i dont get is, how can they build a complete rigging and crowd system in some months but not some basic modeling tools.
There is no complete rigging system. Osiris was their module / proof of concept system that they built. It wasn't full featured and in my opinion a bit clunky to use. There were limitations to it as well. It didn't take a few months from my understanding either. We don't actually know any start / end dates so statements claiming it was only a few months work isn't really fair to say unless we get solid time frames from the Fabric folks.

The Crowd system is something that was years of work. I know this for a fact. Even now it's continuing to get development and evolve. We had to customize a bunch of it for our pipeline and tools btw.
I mean something "simple" like the Softimage "N" Tool to draw polygons. For example for pbp modeling.
Is it fair to make statements like this? I think it can only be said if you know what it takes to develop the feature you're talking about. Many of us I think take things like polygon modeling for granted. If you look at all of the code that the draw polygon tool uses it may surprise us. Luc-Eric any insights? Even a rough estimate of lines of code would be interesting.
Specially with Canvas which would have access to everything inside the application.
Well, you only have access to things that the DCC allows you to. I don't think we could get to the internals of Softimage just because we can us Canvas. What we could do is recreate certain systems that we then have full access to. An example is building a blend shape deformer and blend shape authoring tools. Maybe allowing you to store blend shapes deltas to an alembic file to allow not only the tools but the data to be portable too.
Also a sad side aspect is that tools which companys create will (of course) not be released.
I wonder how fabric will grow then...
Well again, look at what MPC is sharing. It's incredible. The splice stand alone was all them and they were able to share it. Once I get some time I'll be sharing sub-graphs, nodes, and kl extensions. Others are already sharing things in the beta too. I think it is still too early to be thinking negatively and unlike the other companies that have been around for decades, Fabric is still growing along with its community.
Eric Thivierge
Lead Kraken Developer, Fabric Engine
http://fabric-engine.github.io/Kraken

User avatar
blaxxun
Posts: 48
Joined: 04 Mar 2014, 13:23
Location: Austria
Contact:

Re: current Fabric Engine usage?

Post by blaxxun » 10 May 2015, 21:53

No thats totally okay. I even like it, rip it apart.

I didnt mean to sound negative at all. Iam honestly very excited and full of hopes for the future of Fabic.
I just dropped some personal thoughts about this powerful engine.
Im not well informed and also not a programmer.
So from a clueless artists view it looks more easy to write a polygon modeling tool than a rigging system, may it be complete or not.

Iam very sure that it is NOT simple. Thats why i put it in quotes "".

However, it would be wonderful to have a monolitic standalone DCC with the modern core of fabric and with ICE like access to all of its parameters.

Thanks for sheding some light. Very informative.

User avatar
Hirazi Blue
Administrator
Posts: 5107
Joined: 04 Jun 2009, 12:15

Re: current Fabric Engine usage?

Post by Hirazi Blue » 11 May 2015, 11:36

EricTRocks wrote: Well, you only have access to things that the DCC allows you to. I don't think we could get to the internals of Softimage just because we can us Canvas.
Sorry, as I suspect this to be a stupid question:
understanding Fabric is "limited" to what a specific host DCC allows it to do,
how does the "cross-DCC" functionality work?
Is this based on the most restrictive DCC as a norm?
I would imagine Softimage and Modo to expose quite different things for instance...
Stay safe, sane & healthy!

Pooby
Posts: 501
Joined: 27 Aug 2010, 22:25

Re: current Fabric Engine usage?

Post by Pooby » 11 May 2015, 11:58

If you are rigging with Fabric, you'd probably just want to bring in transforms to Fabric, from your Scene control objects and export either transforms to scene bones back, or the final deformation as point positions.
'Spliced' Dccs would have no problem with basics such as that.
When it comes to generating new geometry in the scene, then the DCC needs to be able to accept incoming topology. Lightwave for example isn't designed to have geometry created in Layout, so I suspect Fabric in LW would be a no-go. I would have thought that Modo would be ok though and if not, I'm sure the Foundry will work on areas it needs to make work. Then there's things like UV's Particles etc.

There could be things for example that you might want to pass to and from ICE from Fabric that would make no sense or have no equivalent in another application, but then, if your tool is to be swapped between applications, you'd just avoid designing anything in your tool that is one-app specific.

User avatar
blaxxun
Posts: 48
Joined: 04 Mar 2014, 13:23
Location: Austria
Contact:

Re: current Fabric Engine usage?

Post by blaxxun » 11 May 2015, 14:06

I think Eric missunderstood what i meant with "access to all parameters" of the DCC.
I actually meant building a whole DCC on top of Fabrik.
So then you are actually not bound to the limitations of any existing DCC.

I absolutly understand why the Fabric guys chose this path BUT in the future it would be more than awesome
to have a super powerful modern standalone complete Fabric DCC Suite.

As a Softimage user i really would love to see some light on the horizon for "a bright future"...

Pooby
Posts: 501
Joined: 27 Aug 2010, 22:25

Re: current Fabric Engine usage?

Post by Pooby » 11 May 2015, 14:28

blaxxun wrote:I think Eric missunderstood what i meant with "access to all parameters" of the DCC.
I actually meant building a whole DCC on top of Fabrik.
So then you are actually not bound to the limitations of any existing DCC.

I absolutley understand why the Fabric guys chose this path BUT in the future it would be more than awesome
to have a super powerful modern standalone complete Fabric DCC Suite.

As a Softimage user i really would love to see some light on the horizon for "a bright future"...
There is nothing technically stopping anyone doing that, but the Fabric team are better off spending their time making the framework than making a DCC.
The whole philosophy of Fabric is opposite to the idea of being locked into one DCC, where you have to do things the way the app wants you to. With Fabric, you can make the tools you need, or use ones someone else has made and choose to use them in whichever app you want that supports Fabric, or even make standalone Fabric tools.
No doubt over time, people will make modelling tools, UV tools, etc, and its not inconcievable that someone will try and tie them together into something resembling a DCC.
Whether its labelled a DCC or not, I do like the idea of a set of modules to do my work in the way I want to do it. This is what I do to some extent with ICE anyway.

EricTRocks
Moderator
Posts: 754
Joined: 25 Nov 2009, 01:41
Contact:

Re: current Fabric Engine usage?

Post by EricTRocks » 11 May 2015, 16:58

Hirazi Blue wrote:
EricTRocks wrote: Well, you only have access to things that the DCC allows you to. I don't think we could get to the internals of Softimage just because we can us Canvas.
Sorry, as I suspect this to be a stupid question:
understanding Fabric is "limited" to what a specific host DCC allows it to do,
how does the "cross-DCC" functionality work?
Is this based on the most restrictive DCC as a norm?
I would imagine Softimage and Modo to expose quite different things for instance...
Not allows it to do but allows it to access. Then again it's all about what Fabric exposes into these tools. An example is weight maps. The off the shelf Splice system doesn't allow for hooking into weight maps. We (Ahmidou) have implemented this and rebuilt Splice to have access to weight maps. Maybe there is something in Softimage we can't get to though maybe something to do with the internal Softimage kinematics system. If you can't get to it, you can't use it.
Eric Thivierge
Lead Kraken Developer, Fabric Engine
http://fabric-engine.github.io/Kraken

Post Reply

Who is online

Users browsing this forum: No registered users and 17 guests