When we first played with FE, it was very overwhelming and it was a lot
to take in. They have done tremendous work to make it much easier to
pick-up and run with it with minimum efforts.
Splice made life so much easier and Eric's been creating some pretty
impressive rig set-up very quickly. FE is still young and we have often
faced bugs, but there support is very good and fast so they've sent us
fixes very quickly which allows us to make the tools viable in production.
The first project we tried to use FE on, we ended up pulling the plug in
the end, because we couldn't turn around and completely redesign stuff
as quickly as we needed to. We fell back unto a ICE based solution as it
was much easier to adapt it in the limited amount of time we had. But I
had seen the power of Fabric and I knew that with a bit more learning on
our side and a bit more work on their side, it would become our tool of
choice down the road. It has very much become our spine, even though we
still have a lot of work to do.
We've been working on some tools for a long time so some of them where
built on the old way of doing it and has become extremely complex with
time. But again, they've improved a lot of things which makes it so much
easier to debug, easier to implement new features and transitioning our
tools from one version of FE to the next has become so much easier then
it was on some of the first builds. It depends on what you want to do
really. Some stuff can be very quick and easy, and some stuff is still a
bit long and complex to implement, but it was to be expected knowing
this is still very new technology and very much in fluctuation.
I'm a huge ICE power user so I'm very much awaiting their 2.0 node-based
solution. From what I've been hearing, those who are somewhat
uncomfortable coding will be well served when that comes out. We have a
ton ICE based tools that we will be re-designing in FE once this feature
is released. For that reason, Softimage's demise doesn't really scare me
anymore. It's going to be a lot of work to redo all those tools we've
built over the years, but they will all be so much better with FE behind
them, and they will be easily portable to any DCC we chose to use.
Kind of like ICE was at first, this is a framework so there's not much
out of the box solutions really available. We have to create our own
tools, but once we do, they are impressively fast and extremely
powerful. If you are looking for off the shelf solutions, then going
into something like Houdini is probably the wisest thing to do for now.
There wasn't too many solutions in ICE at the beginning either, and I
think that with time, the community will grow, people will start
creating and sharing or selling tools and it will become even more loved
then ICE ever was.
Working with FE made us realize how old the Softimage infrastructure was
and how much it was holding us back. It was limiting the performance of
our work. I was really impressed by the performance of ICE, but man, we
didn't realize how much more we could get out of our machines until we
started using FE. With FE, there's no intermediate layer that slows our
tools down anymore and we're constantly surprised how much we can pull
out of our workstations. It doesn't have the same comfort of all the
little tools and user friendly workflows that have been accumulated over
the years, but Softimage and other DCCs will always be accessible and FE
was designed in a way that makes it easy to make them work together anyway.
Is FE the solution to all your problems? Of course not, but it's a damn
powerful tool to have when you need to build high-performance custom tools.
-Mathieu
On 28/02/2014 11:40 AM, Eric Thivierge wrote:
Ahem,
In terms of the stuff I've been doing, Fabric is in a really usable
state within Splice right now. Mainly building rigs for me. The
performance is pretty nuts and I've just built a rig that replaces our
old Softimage only setup and I can't give you solid numbers but the
frame rate went from ~4fps to ~20-30fps. Plus it gives us more control
and features.
With the help of Helge to add an additional feature to one of the
built in types, we were able to create the required operator and the
prototyping / implementation was within a week or so. Granted I've
been getting my feet wet with Fabric Engine and KL off and on for 6
months. I've had other things to jump on but if you're a solid TD with
a good amount of programming background it'll only take you a few
weeks to a month to get to a point where you can pump stuff out.
Looking forward to getting off a few other tasks to get back onto the
Splice work.
One cool thing for me is that you can draw OpenGL shapes and
interfaces that allow your animator's to get some better feedback on
what's happening in the rig without having to know any C++. It's just
there in Splice. They are adding some cool stuff in the coming
releases (before 2.0) that's going to be really exciting in many
respects regarding Splice.
If you have any more Q's just post back.
Eric T.
On 2/28/2014 11:21 AM, Paul Doyle wrote:
I would rather guys like Eric and Guillaume (and Mathieu if he has
the time!) commented here - you're going to get a clearer production
perspective imo.