Freeze Transformations For a control in Houdini

[amazingjay]

In maya, one can "freeze transformations" so the controls can be zero-ed out. This means that if the animator wants the controllers to return to the rest position, the animator would just assign a value of [0, 0, 0] to the "translate" parameter for example.

Does anyone know how "freeze transformations" is done in Houdini? (Especially for an auto-rig spine)

Edited December 6, 2010 by amazingjay

[Macha]

Is this what you mean?

I don't know if this works when expressions are involved.

[amazingjay]

Yes, but that "Pre-Transform" option is not available on an Auto-Rig Spine. do you know how I can enable that?

[Macha]

You can right-click any parameter and make its current values default but it won't offset the displayed value to 0 and I don't know of an easy way (expression free) to achieve that. Perhaps somebody else is aware of it?

You can also store poses with rightclick>motion effects.

Edited December 6, 2010 by Macha

[michael]

unlock the spine

make a node

make it the child of the control (example 'rig_ctrl_hip')

turn on Keep Position When Parenting on the null

unparent the null

make it the child of the parent of the control (in this example 'blend__character_scale')

on the parameters of the rig_biped_spine - change Hip Translate to 0, 0, 0

make 'rig_ctrl_hip' the child of the null

avoid putting values in to pre-transforms

[Gaviteros]

unlock the spine

make a node

make it the child of the control (example 'rig_ctrl_hip')

turn on Keep Position When Parenting on the null

unparent the null

make it the child of the parent of the control (in this example 'blend__character_scale')

on the parameters of the rig_biped_spine - change Hip Translate to 0, 0, 0

make 'rig_ctrl_hip' the child of the null

avoid putting values in to pre-transforms

If I may, why is this example better than pre-transforms?

(Sorry to bring up a dead thread.)

In the same scenario we could

unlock the spine and either add the pre-transforms section to the fetch or use the existing one on the merge_blendshapes.

Using the pre-transforms and clearing translates seems like the right idea since we could then use those same tools to extract them later if necessary. As far as I know you can't extract the translation values from the keep position while parenting trick.

Is there some other reason for avoiding pre-transforms?

I admit I have only dabbled in animation and rigging in houdini so I sincerely am unsure.

Edited March 9, 2012 by Gaviteros

[michael]

it isn't...and the fact that an extra node is being added an argument can be made that it will slow the rig down (nulls add up pretty fast) even it if its only by a few ms...

but for a human readable rig, one that you can look at quickly and understand, having the pre-transform on a null of it's own makes things very clear...additionally you're separating the hidden pre-transform values from any parameters that are used by the animator...

what I've done in the past is to build a rig with these extra nulls to hold any pre-transforms, then once I'm ready to release the rig to animation I'll sweep through the network and delete them all - having first set "Keep Position When Parenting" ON on their children - this will push the pre-transform values into their children...but I'll keep a version of the rig with the extra nulls as my working rig and make any new changes there then repeat the deleting when I publish the new rig..

HTH