[Bf-animsys] NLA Object Channels Ordering

Nathan Vegdahl nathanvegdahl at gmail.com
Wed Jun 3 19:44:53 CEST 2009


The idea with the new NLA system, at least as discussed in the IRC
meetings, is that Actions no longer belong to objects, and instead are
their own stand-alone thing (just like objects are).  This is
necessary in order for Actions to be able to affect multiple objects.

It's best conceptualized by thinking of Actions as nothing more than
just stored animations, rather than some weird abstract Blender
concept from ancient days.  With that in mind, animations often (in
fact usually) involve more than just one object.  And those animations
can then be mixed in the NLA editor.

Actions in-and-of-themselves do not affect objects.  They're just
data.  Only through the NLA editor, or action constraints, or
something else that utilizes that data do they affect objects.

To me this makes a great deal more sense than the way 2.4x works,
where actions are somehow specific to an object.

This concept also allows for much simpler and nicer workflow, where we
can have an "active" action that newly created animation data goes
into.  Sort of like edit mode with meshes.  This makes it easy and
clear how to create and edit actions (as long as the UI makes it clear
which action is currently active).

If actions are specific to individual objects, then this easy, simple,
and powerful workflow becomes nearly pointless.

However, there are still a few uses where it is useful for actions to
be linked to specific objects (rigging, for example), so that will
still be allowed.  But my understanding is that that will not be
default.  Perhaps it will be done through a rebasing tool (like
dragging an action to an object in the outliner, links it to the
object and modifies the action's RNA paths as appropriate).

I may be missing something, though.  Aligorith?

--Nathan V


On Wed, Jun 3, 2009 at 6:00 AM, Daniel Salazar - 3Developer.com
<zanqdo at gmail.com> wrote:
> I was testing the progress of the NLA and noticed something odd and
> went to report to Joshua as a bug
>
> http://www.zanqdo.com/tmp/nlastripper.png
>
> in that image what object all those NLA strips belong to? Not "Mesh"
> but a hidden object below all the strips! So as said I went to tell
> Aligorith and he tells me he thinks the same but some people in here
> think that's actually a good idea? I have no clue why but he tells me
> they are following the way the ancient Sequencer works for the
> evaluation order of NLA strips. Well, thats ugly and inconsistent with
> modifier and constraint stacks but its understandable.
>
> Now why the object channel would have to follow this convention that
> makes no sense and will only confuse everyone? It not only makes no
> sense but its absolutely no relevant to the animation result since the
> order of *different* object channels in the NLA editor simply makes no
> difference.
>
> Photoshop adopts this kind of evaluation ordering too, and as I said,
> its ok. Inconsistent with some parts of blender and consistent with
> others but hey its Blender. Now see what Photoshop does in hes layer
> manager with the groups/containers/folders
>
> http://www.binary-artist.com/photoshop_tutorials/images/photoshop_layer.gif
>
> Of course they follow human instinct and don't put the cap on the button!
>
> Daniel Salazar
> Cel: (506) 8366 8384
> Offi: (506) 2219 0511
> www.3developer.com
> _______________________________________________
> Bf-animsys mailing list
> Bf-animsys at blender.org
> http://lists.blender.org/mailman/listinfo/bf-animsys
>



More information about the Bf-animsys mailing list