Houdini 15 Teaser

[Werner]

More to come next week.

[Hirazi Blue]

So what else is new?
(Sorry, I couldn't (edit) resist a slightly cynical remark...)
It's all impressive of course, but I doubt Autodesk, with which most seem to work, to be shaking in their boots.

[Bullit]

Yeah. The simulation didn't even feel right.

[Bullit]

From https://www.facebook.com/particleskull

EXCLUSIVE - Houdini 15 Features

(Probably true, probably exclusive wink emoticon )
-Flip Fluid Lava
-Edge Sliding
-Hougini Engine 2.0
-Layered Material
-Animation POP
-Poly Bridge
-Poly Expand
-Ragdoll
-Renderman
-Retopo Tools

[luchifer]

the new houdini is EPIC.


MauricioPC Moderator Edit: Choosed the correct tag for video.

[MauricioPC]

Looks very exciting to me. That and Modo getting FE Canvas.

[Bullit]

Looks very good, and with tweak tool, if it is like Softimage regarding modeling then it starts to look a full able package.

[Rork]

Holy cr*psnacks! Now THAT'S what I cal an update!!!
It seems somebody took a page from the Softimage workflow and pasted it into Houdini

H15 is going to be an epic release, can't wait to fiddle with it.
Hopefully we get some additions to the Passes/Takes system a la 'renderpasses/partitions/overrides' as well and most of us don't have to look any further for a SI replacement ;-)

But, so much new stuff. It's mind boggling! Love the ragdoll system. "Now with limb detachment!" LOL!!

[Hirazi Blue]

Holy cr*psnacks! Now THAT'S what I cal an update!!!
It seems somebody took a page from the Softimage workflow and pasted it into Houdini

H15 is going to be an epic release, can't wait to fiddle with it.
Hopefully we get some additions to the Passes/Takes system a la 'renderpasses/partitions/overrides' as well and most of us don't have to look any further for a SI replacement ;-)

I fear the moment their innovative approach slows down and all they can offer their eager customers a "measly" viewcube ( still happy with it myself BTW)

[Rork]

But that viewcube will be fully procedural

[SamHowell]

Very cool. A very solid and well rounded release. They seem to be aiming more for the generalist and small studio market (the natural home of Softies) as well as FX. It makes me a lot more confident in my decision to pick Houdini for my future.

I could be wrong but I'm pretty sure a lot of those new features were requests made by the users on the Sidefx forum. The modelling tools, material library, faster FEM (cloth) and character animation were all mentioned on the discussion threads. This show's a healthy relationship with the Houdini userbase.

Good times.

[Mathaeus]

I fear the moment their innovative approach slows down and all they can offer their eager customers a "measly" viewcube ( still happy with it myself BTW)

Don't need to worry, as they already have a plenty of room to "innovate", just to reach the level of standard 3d app I mean, if that "faster FEM cloth" means even twice faster, great. But, that I know, even old, single threaded SI Syflex shows around 5 - 7 times better speed, so... there is a space for a last two "improvements" of H cloth.
Similar goes for kinematics - an insane mix of unbelievable basic set of direct tools (has no even self - blending constraints, rotation order is just fixed in some cases) - and to hell abstract CHOPs for anything more. Looks like a developer's prototype of kinematics for 3d Studio Max 1, occasionally leaked into public. I've intentionally mentioned Max, as this app already provides really strong and usable processing over animation channels - while all that per - channel detailing doesn't help when it comes to a lot of them, like with characters.

Also, who knows how all that "new" will show in reality - for example, heavily advertised Open subdiv in v 13, wasn't able to interpolate point attributes, in some cases, during all v 13 period - so, something to generally avoid in SOP network...

Will see, I really hope I'm just pessimist, but "Houdini as usable main 3d app for generalist", sounds like reality of H v 20 or something - and only for those who can't live without unix wording everywhere (@ ..../.... $ and such historical crap). Most likely, will be a standard H release: one new solver, one or two disastrous 'direct' implementations like hair brushing tools in v 14, a lot of small enhancements over 3d app which is already out of criteria. That's H : firstly you're forced to walk backwards, then there's team to generously provide the retro visors.

Where's that Fabric Canvas.....

[Hirazi Blue]

But as we all know: universal frameworks (Fabric, more or less) only work well if there is competition around...

[Mathaeus]

But as we all know: universal frameworks (Fabric, more or less) only work well if there is competition around...

While I have to admit that I've tried to avoid the external thing like Fabric, now I think the main advantage is possibility of 'limited framework' like ICE or Fabric. "Limited" because it's possible to use it only where you want in your 3d app, not full time.

By the way, one interesting thing when it comes to H, is really broad range of 'what it can do', but very very strict rules of 'how it can do'. Basically, each SOP node acts like separate ICE tree, taking the result and attributes, that's it. While inside one node, only one 'turn' is possible - it can't take data from previous 'row' and do something again with this. In reality this means, it's close to impossible to re - assemble the functionality of typical ICE 'logical bricks' system, like SPH addon, Strand Tree, Motion Tools and so on, full of intermediate, modulating nodes. At least, not only by factory H nodes, without forcing the user to manually type (?!) the modulated attribute name in some SOP, somewhere far earlier in tree. Or, without additional network of expressions all around, which is not only convoluted (imho) way - it also destroys the flexibility, down to "Poser style" of tons of sliders, nothing else (for artist).

While in Fabric, it seems to be perfectly possible to get the 'ICE way'.

Just to stop with long stories here, definitively there is a feeling of one familiar (for us) 'framework' in H. Famous 'can do everything, can not compete against particular anyone, every next action is overcomplicated'. But, Mental Ray was an industry standard, while H is not. That is, because I've used H only in 'pleasure' category (never felt enough confident for 'business'), no big deal to switch to Fabric.

Anyway, found H nice for scene assembly, Mantra is really nice for outdoor stills, as long as there's enough of stuff created somewhere else.