While Houdini may be technically more powerful,
ICE, even if still can be considered somewhat technical, is quite known
to be much more approachable and is (yet) another testament of bringing
complexity to easier reach.
Houdini seems to be mid-way between actually scripting, and "higher
level" visual programming.
We will wait and see how bifrost will be on the friendly (and
integrated) side
(after foam and everything else concerning naiad will be implemented..
and graph actually exposed)
Here is a known (performance only) comparison between ICE and VOP
http://frenchdog.wordpress.com/2009/09/12/ice-vs-vop/
And here is a typical and quite fair view of SI & Houdini
|
|
|
|
If
you
want
a complete package (modeling, animation, procedural stuff, VFX
stuff and rendering), Softimage is easily ahead of Houdini.
This
is because Softimage started out as a "regular" CG suite, like Max and
Maya. All the everyday workflow is very efficient and convenient. Of
course you also get some killer features like FaceRobot with auto lip
sync, GATOR, MOTOR, Lagoa, Gigapoly, Animation mixer, history stacks
etc.
And
then if you want to do procedural stuff, you can do that too. ICE has
matured nicely and can make+deform geometry, do particle sims, make
control rigs, affect any attributes in the scene including SRT, all
without needing to code or write expressions (in Houdini you write a
lot of expressions... stamp("../Pathsareannoying/copy2"), $CY,
$DoIHaveToRememberthis?, 0, 0, 1)
Houdini
still has more raw power (except speed, because multithreaded ICE is
very fast). If you're highly technical, know how to program and you're
mainly interested in VFX then Houdini is probably still your best
choice. But if you want a complete package, Softimage is preferrable.
Especially once Arnold is available to the general public, which
shouldn't be long now. Arnold is killer. It obliterates every other
renderer.
|
>From Pooby
-
.
.
. ICE could be even more and it needs some competition.
(Houdini's approach is currently far more reliant on knowledge of maths
and coding so I dont see it as a similar beast, and houdini isn't a
geometry animation centric package either like Softimage is)
sdsd
| Mathaeus |
|
Post
subject: Re: ICE vs
HOUDINI - Split test results (a comparison)
Posted: 04 Sep 2013, 09:15
|
|
|
|
I'd
also
believe
test is 'valid' today, too. At least when it comes to part
where ICE is faster, that is, 'point pushing'. Actually, I think ICE
can be even much much more faster when it comes to strands, as
specially optimized, 'child' of particles.
There is price for strands performance, as usually, some geo query
won't catch the strand.
As far as I know, something similar to ICE strand doesn't exist in
Houdini.
I think Houdini also does not have ready-made nodes like ICE Get
Closest Location (working horse for so many tasks).
But, I think differences are result of intentional, design decision.
Generally Houdini looks like something more about flexibility, instead
of performance, while ICE is more artistic, optimized tool.
Artistic, here = (relatively) limited set of tools, together with
(relatively) high level of user protection, allowing the creativity as
much possible.
Strong set of 'classic' operator in SI, makes ICE
even more artistic - you're allowed to use ICE where, if , you want.
This is not possible in Houdini.
BTW, my knowledge about Houdini belongs to playing with demo, just a
bit more this summer than usually, that's all.
|
|
On 03/21/14 5:56, Andy Goehler wrote:
Sebastian, take it slowly and all will be well ;-)
As much as people like to refer VOPs/VEX to ICE, I’d say keep in
mind they are quite different.
While ICE cracks tend to go hardcore in ‘one' Tree, the Houdini
guys use VOPs in conjunction with the other OPs available to make up
the overall result.
See you over at the SESI forum.
Andy
|
