I was watching this video for the Maya game animation workflow in Destiny by Richard Lico
https://www.youtube.com/watch?v=8p0WQJtjZZ0 [youtube.com]

basically it's dynamically building a rig and then working with baked animation data
and then changing the pivot space for certain rigs to clean up the animation keys

I was wondering if anyone has the experience to tell whether or not this animation style could be replicated in houdini,
thanks

Can you clarify what you mean by “dynamically building a rig”? I keep looking for it but I can't find it in the clip.

Also, I can't find the “changing the pivot space for certain rigs to cleanup animation keys”


This is what I saw, a semi-automated rig (moving the head moves the torso a bit to save time animating) and at the end of the clip I saw he building constraints to offset animation and overlap animation.

Is this what you meant?

Hi guys,

I think he has a set of tools going for being able to take a set of baked out channels (for want of a better word) on the joints and stick a rig into the middle of that existing animation, which he can then use to modify the movement before re-applying the channels onto the joints. The idea seems to be that you only ever have keys set on the joints themselves… I have to admit, I quite like the concept.

Keying any IK/FK blending always seems a bit dirty to me, I think I would invariably end up using the IK handles to get a pose and then matching the FK to that position, always keeping the keyframes on the FK channels. I think that's closely related to the approach that Mr Lico is taking.

The changing pivot space still confuses me a bit though… maybe you're referring to the use of locators to pass around animation data? i.e. to create offsets and lag in the movement between hand and shoulder

ah sorry!
thought there'd be some workflow in that video

https://vimeo.com/171165613 [vimeo.com]
here richard shows his pipeline to output a skinned character into a game engine from maya.
Friedasparagus seems to get it.
Richard destroys the rig, then rebuild a rig over the same skinned character. in this method, he can rebuild the rig in any curve space he wants for what goal he's trying to fix.

https://vimeo.com/171164594 [vimeo.com]
in here he shows his methods to get some quick animation overlaps using temporary aim constraints

the methodology is using the best rig as a tool for what goal you're trying to do.
it's a really unique approach compared to the try to build a rig that does it all approach.

Moin,

finally having watched through the videos, I do think that recreating this workflow in Houdini should be possible without too much hassle - after all, it isn't really automatic, but depends on scripts that create rig-controls for specific characters. Meaning: The scripts are “character-dependent” (or at least “character-base-rig-dependent”).

The bit that confused me the most was that - to me - a rig is ANYTHING controlling something. A set of bones that deform the geometry already *is* *a* *rig*. In Mr. Lico's terminology, the rig only is the animator's control layer topmost, so the bone-structure doesn't change at all, it's “baked” into the character. Only the controls (“rig”) get added/changed/modified dependent on what the animator wants to do.

And this definitely is doable in any DCC. I somehow doubt that this workflow is that helpful if the geometry changes, because obviously the weighting (“skinning”, “capturing”) isn't redone when the “rig” (Lico-terminology, i.e. the animator-control-layer) gets deleted or changed. What I consider the “basic rig”, the “rig-rig”, if you will, remains unaltered. I do know workflows where iterations between animation, rigging and modelling would suffer from the constant need to adjust the bake-procedure and, very likely, the script-side of this.
But if your characters are “set in stone” and you know that the weighting doesn't change, then the idea of “hard” separating the animation controls from the functional part of the rig seems very mean and clean

Marc

ahh thanks for the input Malbrecht,

I see what you mean by defining the skeleton with the binded weights as a rig itself. i think the benefits are very clear still. http://www.gdcvault.com/play/1020441/Designing-the-Bungie-Animation [gdcvault.com] at 20 min he shows the tools they used for the animation cleanup/ balance/ parent space of the rig.

and as long as it's planned out right, new geo on something doesn't seem like it would break this method. http://www.gdcvault.com/play/1020441/Designing-the-Bungie-Animation [gdcvault.com] . In the same destiny video GDC talk with Richard they were showing animations from smaller characters being shared to larger ones with an added delta offset animation layer on a custom runtime animation rig that's very compressed (he mentions 500kb to 20kb), at around 41min in . it was a really cool talk with what tools they're using at bungie for the destiny animation pipe.


how would one approach this in houdini? is there an output window for actions like maya's mel that I could use as reference to build out scripted actions in houdini? and what script language would it use?

Hi,

I am in the midth of a video production, so I can't look into the other videos, but the topic definitely is of interest to me, I love learning new tricks

What I mean by “back and forth with animation, rigging and modelling” probably does not apply that much to game characters where the overall amount of geometry is limited plus some “common sense rules” apply to rig/geometry dependencies (like “three edge loops on elbows” - just an example, obviously).

There is a video tutorial about auto-creating weight maps based on (properly named) geometry areas / face groups. I do think that this could be a good brainstorming start for creating something like Mr. Lico's system. Some kind of autorig-and-auto-weight setup. For example: Constraints work in Houdini, so if your character has been “pre-rigged” (if you forgive the sloppy terminology) there is no reason why reproducing that “ear-aim-at-locator” shouldn't work out of the box without scripting.
Scripting only comes into play if you have to adjust or even create the prerig/skeleton or have to prepare (hinted at above) specific “standard functions” (e.g. adding an IK chain to legs, probably using a Python function/HDA that looks for bones / joints named “Leg_1” to “Foot_3” or whatever and creates the “default IK chain”). Because IK does not get stored into FBX, as it depends on the DCC's implementation of a solver.

For more complex rigs I am almost certain that this method may have *some* benefits but at the cost of some overhead that you may not want. For “simple” rigs it still looks neat.

As always, just my 3.1415andsoon cents.

Marc

Hey Sirvoxelot,
At the moment I can't be of great help but wanted to mention I am chasing similar techniques in houdini. Being very used to motionbuilder I already got pleasantly acquainted to plotting to rigs on the fly and doing this on an even more custom and modular basis would be golden.
Earlier I got also attracted to richard loco's methods for inserting dynamic secondary motions now he is working at polyarc.
I am investigating this on the moment by methods of motioneffects (lag) and/or wiresolvers.
After this, creative and destructive rig-baking are on my list.
In my view there is no distinction between game animation or film animation. These methods are very liberating and inspiring for everyone who is not in the search for the perfect rig and the most polished graph curve.
I'd gladly keep you posted and am interested in your searches..

yeah,
also I found this
https://www.mokastudio.com/features.html [www.mokastudio.com]
haven't got it to work on my 4k monitors at work


still need to test it at home
being able to sketch poses looks awesome. and being able to dynamically change the ik to pose what you want
looks to be really innovating


keep me posted !

so I found another interesting method of animation that a fellow named Raf Anzovin is using
he calls it ephemeral rigging, the results are an interpolationless animation that looks almost hand-drawn
animation pilot the new pioneers [vimeo.com]

there's more details in his blog post
https://www.justtodosomethingbad.com/blog/2017/9/17/its-a-post [www.justtodosomethingbad.com]


he explains some node callback function in maya is what he uses
https://www.justtodosomethingbad.com/blog/2017/10/22/nuts-and-bolts [www.justtodosomethingbad.com]

would something like this work well in houdini?