Mecanim : Driven Additive Sweeps

In a previous life, to make a character aim a gun, we have used a couple of driven additive sweep animations, masked on their upper body.

I can't see a way of driving an animation's playback position in Mecanim.

To be a little more clear on exactly what we were doing, we would have a base run with gun animation. We would also have an additive sweep gun from left to right animation and an additive sweep gun from upper to lower. Then we would drive the play time of both the additive animations to make the character point the gun anywhere within the circle formed by the left/right up/down sweeps.

This video shows the technique (jump to 40 seconds in)

 

I too am interested in how this is possible with mecanim!!

 

FK, our animators are used to authoring them this way.

Have exactly that. The problem is the 'driven' part. There is no way to programatically scrub the play position of the additive animations in the current api. Unless I missed something?

Thanks

Dan

 

@DanTreble,

In Unity 4 beta 11 you have now the function animator.ForceNormalizedTime(float normalizedTime) so that you can set the exact normalized time of the current state or transition.
Is it what you call "play position"?

Regards,

Sam

 

Thanks Sam, how do I target animator.ForceStateNormalizedTime at a specific layer?

 

Hmm, I didn't know that was public.

That's actually a method that we used internally, but it was deemed safe to make public. It's in the process of being renamed to ForceStateNormalizedTime and it only works on the base layer!

So you cannot do a horizontal and vertical sweep as the OP wants. But perhaps he can settle with either horizontal or vertical sweep and then use a blend tree for the other axis? Be advised that this functionality is likely to change in a later 4.x release. We do want to support driven additive sweeps in later releases but its just too late to add to 4.0.

EDIT: I just tested it with additional layers and transitions to be entirely sure how everything works. You must call it every frame and if there is a transition in layer 0, both the old and new state time will be forced. The transition is completely unaffected by the forced time (except transitions using exit time because exit time needs to be exact match when time is forced). All other layers are completely unaffected.