So it's actually the riggers that are missing in the first place
I had chats with quite a number of riggers at various studios and somehow the exchange was pretty similar: “Oh, you're a Houdini guy! I really want to get into Houdini, I hear it's very flexible etc etc”
Being technical, riggers are immediately attracted by Houdini. Yet, somehow, they don't seem to make the jump.
A pro level rigger versed in Houdini is something I could use at my studio as well

Exactly!
Even if I wanted to make a switch and use H for character animation as well, I'm missing rigger .. as a lot of others I assume

Mirko Jankovic
If it is possible to jump into it, have some nice rigs and just start animating and using it, it would be a start at least. Checking out animation tools and what is missing, giving some feedback hoping to get them more polished (and hopefully close to Softimage level)

That being said, I still don't get what's wrong with the included rigs and autorig toolset, to do exactly what you're saying above?
And if you feel there's something wrong with those rigs/tools, you could use your exploring to feedback on those. SESI are listening.

imho, riggers will come once more animators will like to work in Houdini. As I've seen many times before, riggers love working in Houdini, there might be some glitches, but on sight people feel how beneficial this environment is for their work.

The problem is that practically speaking the transition is much more feasible for freelancers than even small studios. For example, most europeans feature animation projects split work between companies these days. Maya has became a backbone of shared animation pipeline (for rigs mostly). This is also the case for game cinematics, where you often get meshes and rigs from previous vendor. Unfortunately this situation puts hard constrains on pipeline decisions.

symek
imho, riggers will come once more animators will like to work in Houdini.

This statement echos my stance on the “chicken&egg” issue of who makes what, which is that you have to first make your product good enough to be desired.
Some people have a different position - “you create the product when you have demand” and while this is true for untried market segments in which you'd be taking a risk by creating something you don't know you're going to sell, in our case, there is already a demand and currently it is satisfied by another product from another company.
(talking about character animation here)

Ok so we basically figured out or established state of character animation in Houdini.
Next two things are.. try to locate more Houdini riggers as one, and another one see what can be done with current things in H for animation and what is missing

Just started some testing in between work with provided character and 3 things that noticed right away, and will be adding more hopefully sorting out any slowers or stoppers along the way.

1. FPS rate in view port, it wasn't really that intuitive at first, bin looking and looking.. wasn't really intuitive that turning on FPS notification in view port is actually Display Options > Additional Informationhow Time
But no biog deal, once figured guess can remember that heheh

2. Is there any source of some useful animators scripts around? I usually don't like using them but over the time they can be huge time savers, to start for example an mirror pose script would be great. Also I've seen some pose manager in there but haven't tested it yet so need to dig into that but nice pose and animation library manager would be great. Anything like that around? I really never used too many scripts back in Softimage.. it provided most of needed stuff out of the box.
In maya it is whole new thing and without scripts you are crippled…
Hope that Houdini will be more like Softimage in that way and that most of needed tools will be provided and properly implemented and no need for 3rd party stuff. Also most of these could be already there but as only now starting .. part of the fun is locating things, hard to find something if you don;t know if it is there at first place

3. are there options in Houdini to have built sort of GUI picker like Softimage's synoptic view or picker like the one from anim school, makes selection of controls easy and provides couple other options and helping tools.

Just a small thing to start after testing first provided rig a bit. And see how it goes…

http://www.sidefx.com/docs/houdini/pypanel/_index [www.sidefx.com]

http://www.sidefx.com/docs/houdini/ref/panes/pythonpanel [www.sidefx.com]

this whole thread screams out a need for the Alembic of rigs - compatibility and exchange format. Who is working on one and when will it be available is the only question needed here.

Well hows that Kraken from Fabric shaping up, what is going on with that? didn't get too much traction?

https://vimeo.com/103180191 [vimeo.com]

ya - it's just so dopey that even for fx work to move just one joint the file has to be tweaked in Maya.

digitallysane
That being said, I still don't get what's wrong with the included rigs and autorig toolset, to do exactly what you're saying above?
And if you feel there's something wrong with those rigs/tools, you could use your exploring to feedback on those. SESI are listening.

They are looking like great effort to keep the realtime playback, which is good move, compared to previous rigs in Houdini, probably aimed to show some possibilities. However, structure is pretty much simplified (to keep the speed, probably), for example I don't see independent deformer carrier hierarchy (shadow rig in Softimage world) which makes them problematic for exporting and many other tasks. Also I found cache node in some of them, at end of deformed geo(so, full speed is possible only after initial pre-roll), which could be considered as doping in rigging world, when it comes to such simple structures.
Even with all that, playback is not great, which imho tells what is main problem in H : still not existent chicken. Too slow playback, that's it. And Side Effects won't be able to change that with current architecture I'm afraid, because they already spent the all dirty tricks with these rigs.
For comparison, in Maya is not big deal to have literally everything constantly alive (guides, controllers, initial hierarchy, transform processing, several deformer carriers), having something conceptually similar to modern general purpose shader, where is matter of interface to have a certain set of controllers exposed. All that with playback that can't exists in H (H limit is 120 fps if I'm correct), and completely non destructive. H old school ‘creation of rig can’t be undone' message sounds archaic, and sooo destructive.
Plus somehow standard set of missing basic features in H. When I become to H board with H 13, noticed that H is only 3d on planet having a fixed bone rotation order (something probably created for specific case of specific mocap import, according to z-something rotation order) …. and still is, in 2017. And so on and so on…..

Anyway there's still significant area of auxiliary, not real time (today) tasks where H should be nice: muscles, crowds, for small example I'm slowly working on procedural weighting system in H for use with big counts of deformers in Maya, for cases where any general method (biharmonic or such). However, for ‘core’ work, I don't think so.

Mirko Jankovic
Well hows that Kraken from Fabric shaping up, what is going on with that? didn't get too much traction?

https://vimeo.com/103180191 [vimeo.com]

Imho Fabric has a nice points over H, being able to work directly in Maya. Structure is a way more versatile (imho), able to output the same network directly to points, transforms, whatever. But, well, it's external engine, imho it has sense for something really really complex. Also it's not easy to keep the decent speed, well at least not with my knowledge, especially in Maya Parallel era. While networks are easier and more logical than in H (imho), nodes are a way more technical, things like safe and ‘unsafe’ matrix invert, C++ naming, so on.

amm
Even with all that, playback is not great, which imho tells what is main problem in H : still not existent chicken. Too slow playback, that's it. And Side Effects won't be able to change that with current architecture I'm afraid, because they already spent the all dirty tricks with these rigs.
For comparison, in Maya is not big deal to have literally everything constantly alive (guides, controllers, initial hierarchy, transform processing, several deformer carriers), having something conceptually similar to modern general purpose shader, where is matter of interface to have a certain set of controllers exposed.

By this criteria, nothing should ever be done, in any area.
One can get back in time when DOPs made their appearance. Various comparisons with established (or emerging) solutions (Real Flow, Naiad) appeared, and in lots of areas, they were not necessarily favorable. Still, there were advantages (the architecture was superior, flexibility etc) and a constant evolution.
You have to start somewhere and right now Houdini is a competent character animation solution that is evolving. Yes, it still has to play catch-up in many areas when compared with the established solutions, but it also offers some big advantages, especially for small studios and teams.
And I agree with some of the stuff from Fabric, I do hope (and I think it will happen) that SESI should move some parts of the rigging stuff in SOPs and make it more generic.

With inevitable “I hate to say that”, when it comes to rigging and animation, and I'm afraid in a lot of other fields, Houdini is simply not able to set criteria. Not an option. It has to follow, just like Blender is doing for example. One day…. when it will have animator user base comparable to something like Blender, it will be possible to negotiate about some option, in best case. While yeah, to reach that level, it has to follow……
Or, it can to do businesses as usual, trying to fill the gaps of big players, as an complement.

Pretty much what I was saying.
Not clear what “In lot of other fields” might mean.

Well honestly I do hope that Houdini will keep progress like seen in last versions ie, improving modeling and UI user experience, and to spill that on animation and other areas.. really after Softimage is killed a lot of people are still looking for nice similar all in one solution. Houdini is known of t's steep learning curve which is being kinda helped out lately and filling in gaps hope to get to a nice generalist tool as well and not only VFX go to solution.
But also all other areas have many different options, modeling got ton of misc solution, from all standard maya, max, blender, modo to sculpting and other retopology options. Even shading and rendering is in there but character animation feels so neglected that Softimage and Maya, and now only maya are kinda only viable customer ready solutions. And maya is known crap to anyone that used Softimage so we are left without any choice really. Maya if you wanna be relevant as character animation freelance pool, or Softimage for some personal projects or stuff where you are working on your own and not in maya centric studio. Max is complete crap in that field and others .. even more. Couple animation only projects failed miserably.. like that what was it.. nakayama nok.. something like that..

So hopefully Houdini will be able to fill the gap.

@amm so I'm not sure why you keep trying Houdini then? I appreciate the critiques but it would be cool hear the context of why you stay with Houdini. Thx!

I'm not on Houdini yet, I'm looking into option of switching to it more and learning it more as well.
Softimage, unfortunately is on it;s way out.. workflow wise and a lot of stuff in there it will still be best option for another couple years but jobs looking for SI guys are dropping fast, finding SI people is harder and harder and tech is simply going over it… to start with missing 4k scaling support and gap will just go bigger and bigger.
Not to mention that all other developer of 3rd party stuff is just ignoring SI so for example getting stuff from SI to some game engines.well most of them is problem already. So even having SI covering most of stuf now better then others.. time goes fast…

Maya for small team or single guy generalist is big NO, so retarded workflow is hard to imagine…
Max is crappy plugin for autocad, Houdini is already better in character animation then Max
Blender is opensource mess… bunch of probably good things tied loosely together in an inconsistent UI…

Do I need to add more.
Simple as times goes buy SI will be phased out. From simple character animation point Maya is unfortunately here to stay, but hopefully new kid on the block will start changing that.

I mean it is just fst rumbling of ton of stuff it is hard to just put in something that is not a wall of text or an evening of beer and chat but…

At the end.. why not learning another software that proves to be progressing. Yes, there is less and less time to learn new stuff, and other reasons but .. why not

It wasn't meant to be critique of Houdini character animation state but more of exploring where is it now, what is missing and is it on the way to become viable character animation package.

Mirko Jankovic
workflow wise and a lot of stuff in there it will still be best option for another couple years

Which shows why these chats are always so complicated.
While you think the above, IMO Houdini was always a superior overall package to XSI (and a great source of inspiration, especially for ICE), with a much better fundamental approach and workflow. Yes, in some key areas XSI was much better (interactive modelling and char anim), but overall it felt to me like an inept mess with a nice shell (and an over enthusiastic marketing team). I guess personal history plays a big role in all these discussions, that's why I don't take these arguments so seriously.

Mirko Jankovic
It wasn't meant to be critique of Houdini character animation state but more of exploring where is it now, what is missing and is it on the way to become viable character animation package.

I can only repeat what I was saying: it's already a competent (even if not a “leading”) character package, that is improving fast. It also has workflow and architectural features that help enormously with character projects.
And the company is listening, so you have a much bigger chance to get involved in the development of the character tools and workflows (it's not clear to me that people realise how important this is).

Workflow wise it takes some time to get used to.
For example I've started first on 3d studio (dos version, not even max), after some 3dSMax jump into Maya 2.5 and that was already huge improvement and felt much better workflow wise. And then again going from Maya 8 to Softimage 7 was again huge improvement, but over all it was like using what you are used too but improved version. But overall it was close enough to make fast jump.
With Houdini it is not really that simple as there is a bit bigger difference in whole logic so it is taking a bit more time. There are strength in Houdini but being so different from others is what makes jump a bit slower. Just slower not impossible.
On the other hand, Blender that for who knows what reason introduces completely nonstandard view port interaction for example was harder to me then Houdini You go to somewhat familiar interface with viewport and all and you can;t start even going around the scene in somewhat logical and what is used to be way in all major applications… Really bad. For me at least. In Houdini at least fundamental things are same and then just need to build around new logic.

As said like a while ago.. for exmaple:
In Maya you workaround,
In Softimage you work.
And yes SI feels like a glow. But again that could be because of getting used to but also having really nice organised workflow.
I'm still lacking time spent in Houdini to be able to figure out if it is workaround or work model
Hope it to be second one. Again I do not NEED to learn Houdini nor having some immediate pressure but for future proofing it seems to be on a better path then Maya and honestly every day having to use maya for work is PAIN!
So why not try something else just to keep sane.

Back on Houdini and character animation, you may be perfectly right BUT there is still lack of pool of riggers. Let's say I have a project and need nice rig done in Houdini. Where to find rigger?
As a matter of fact I did had project and even twas thinking to use it for testing animation in Houdini as it is rather simple but riggers.. no go. So back to Softimage.

I could use this to call out for any Houdini riggers as well to add to my page www.cgfolio.com
Just something I created a while ago to kinda have nice pool of people and be able to filter out fast the skill needed and availability and get in contact with talents needed faster then standard post ad, wait for reply, go through emails… contact them .. etc.. was looking days if not weeks searching for people like that.
So any riggers and other skillet around with Houdini talents are welcomed to add to the list