Constrain object to deformed curve?

[strasser]

man!!! this is driving me nuts.....
is there any way to constrain a cluster to a curve, which is animated AND deformed by a bend deformer? i have it follow the animation but NOT the deformation. i tried everything i could think of and searched the web but i couldn

[joojaa]

If you want to constrain something to a curve then path animate the custera long teh curve but insttead of changing the along curve values detach the anim curve and just leave a static value on the animation parametter for example. Jut one of 3 ways to do it offcourse.

[strasser]

hi joojaa,
i see you really are a full time poster and a very helpful guy as well, thanks for the reply. if i

[joojaa]

Aaah yeah... well in taht case it sounds like a fdouble trransform of come kind! Im a fraid that i might have mis understood your problem a bit (the other ways are to use a curve ).

Is the cluster part of the curves controll structure or is it justa clouter controlling something NOT in the curves past history?

Woud also be interested in the hierarchy youve set up!

[joojaa]

OH yeah now i can understand the problem fully... And its tricky!

The cluster isnt part of the deforming surface... It would be better if it was... You se you could detach the curves points from the deformer and just attach the cluster to the defrmer instead. This however doesent yield 100% acceptabe in cases where the cluster binds more than 1 point!

Im timing the most optimal solution just now....

[strasser]

okay,
the setup is as follows:
the curve represents a fishing line and is part of the splineIK setup for it. the curve is pointconstraint to the rod, which is animated and deformed by a bend deformer, the first cv of the curve is also part of the bendset, so it stays connected to the tip of the rod. i have one cluster for each cv of the curve (except the first one), the clusters are grouped to themself and the pivots of the cluster groups are snapped to the tip of the rod, so i can rotate them around the tip to make the line swing and look as i need it.
the matter is i need the clusters and their groups to stay at the curve

[joojaa]

create adummy group contsrained to the end in some way! Then make a relative cluster place that pivot to the end of your rod... Now this method has a drawback the cluseters and the manipulats dont reside where the point aculay is but works like you want anyway (teh ui jut doesent follow).

So if you realy realy need the ui to be clean AS well then you need to use a non relative cluster and ucse connection editor to read the worldspace poitions of the group greated to its rotate pivot... And the handles position.

[joojaa]

create adummy group contsrained to the end in some way! Then make a relative cluster place that pivot to the end of your rod... Now this method has a drawback the cluseters and the manipulats dont reside where the point aculay is but works like you want anyway (teh ui jut doesent follow).

So if you realy realy need the ui to be clean AS well then you need to use a non relative cluster and ucse connection editor to read the worldspace poitions of the group greated to its rotate pivot... And the handles position.

[joojaa]

create adummy group contsrained to the end in some way! Then make a relative cluster place that pivot to the end of your rod... Now this method has a drawback the cluseters and the manipulats dont reside where the point aculay is but works like you want anyway (teh ui jut doesent follow).

So if you realy realy need the ui to be clean AS well then you need to use a non relative cluster and ucse connection editor to read the worldspace poitions of the group greated to its rotate pivot... And the handles position.

[failure]

It seems that you have an infinite loop in your scene graph. You want curve to drive the clusters and clusters to drive the curve.
So this idea is not going to work clean whatever you use for constrainment. Can you get away with animating cv's directly? ( Put the curve pivot at the tip )

[strasser]

oh ok, sounds interesting
can a dummy group (a null empty group, i suppose) be part of a deformer

[strasser]

hi failure,
you and joojaa had alot work with me the last week, you replied to all of my questions, thanks for that.
yep that loop thingy occured to me as well, but i wasn

[failure]

Hi,
I'm sure it can be hacked to work somewhere along the line of joojaa's suggestion.
I'm also pretty sure there's a "logical" flaw in your setup causing the circularity in the graph.
So try to rethink the setup and maybe find different / simpler solutions to your deformer architecture and dependancy

[joojaa]

Well it can but the more i think about it the more it seems that the setup contains sevaal redundant layers of controll...

I mean if the idea behind the clusters are to enable you bend the curve more easily then why hae a spline controller at all i mean the ik chain contains (the ik chain would even be tyetter in many ways) all the neccesery things for the animation anyway... With much less overhead.

[failure]

Agree.
Think it's a good idea to get rid of the clusters.

[strasser]

i see, maybe i wasn