Cineversity

Trying to get a belt to animate with a nice rounded move.  I am attempting to use MoSpline with a Spline Wrap Effector.

I’ve got one side curling nicely.  But I can’t seem to get the other side to work with a second Spline wrap and MoSpline.

The first one requires an offset off -100 to get the curl to start from the center.

Is this a bad approach?  I am thinking maybe I need to use bones and connect the bones to the MoSpline.  Would that be a better approach?

Not quite sure how to connect the bones without a tedious Constraint at every joint approach.  Then I will need to weight it properly.

I was hoping I could get the spline wrap to work.  Any thoughts?

Thanks

TAK

Attaching a jpg

Hi TAK,

So far I can read about your target, two Bend-Deformers would do the trick. If not, please provide more details. :o) I will have a look into further options, if.

Happy New Year

Thanks Doc.

I ended up using bend deformers but for future animations I may need a little more control.  So I wanted to use Bones/Joints.  The problem was getting them to curl smoothly.  I recall some sort of Expression that made each bone rotate at some ratio to the previous bone… 

I’ve also been trying to get more familiar with MoSpline so I was trying to work out the “Double” Mospline/Spline Wrap concept… which still doesn’t quite work.  But MoSpline has the Twist, Bend, Curl all built into one which is nice.  I played around with restriction tags but that didn’t quite work.

So the Bend Deformers got me through.

Happy New Year

Nice that you found something that worked for you, TAK.

The Idea with the curled Joints: it is based on an cascading information. Sounds cryptic, but it is simple.

Create a long chain with Joints (Manually with the joint tool or with the “Spline to Joint” command.)

Bind the joints to the object you like to deform.

Select all the joints in question and in the “R” field that you like to use, type in a number and click on the animation dot (R16).
If the “curling” is supposed to be more progressive, use the “num*3” (3 is a suggestion, use yours!) - record.

Done. Perhaps, use a SubDivision Surface [HyperNURBS] :o)

Check the example, which was done in a few minutes, including testing, as I do normally—everything is subject to change in life! :o)

The “…11.c4d” file shows some variations. Note that I adjusted the angle between the two middle joints, as they would otherwise double the amount…

Enjoy.

Thanks doc…  The idea of keeping the curling accurate.  If it is a simple curl then all the R angles are the same. 

For the other curls you seem to have offset each one by a specific amount….  And the difference between each angle is the same… let’s say multiples of 3.  3, 6, 9, 12

I suspect you could set up an Xpresso tag with a slider control to control the difference between each angle.  I think that will be my test for the day.

Thanks !

Hi TAK,

You’re welcome.

XPresso might be a way, but the examples are done as mentioned above “num*value”  (num for the position in the selection—Object Manager based, and value for the initial angel).

Too really get the point across, I even placed an image to the post, have a look.

I have even placed an idea in it how to create an S shape—at once—in that way. It is in the files. :o)

All the best

I guess I am referring specifically to the sort of Helical curl.  Each joint’s value has an identical differential to the previous joint.  Did you manually plug in those numbers?  An expresso would allow you to input different variables.

There was a guy at NAB a few years back who did a whole talk on the Fibonacci sequence and how to control curves.  I will have to find that again.

THANKS

OK. I will try it in an other way

Joint.0
0* value

Joint.1
1* value

Joint.2
2* value

Joint.3
3* value

for four joints.

If you select the joints—all four, you can also type in num*value

========

Fibonacci, the sum of the two before equals the next one. http://en.wikipedia.org/wiki/Fibonacci_number

I would go with (num*6765/4181)*value [to get close to it, simplified], but the hierarchy might be in the way, as it cascades the values as well. My suggestion would be (num+1)*value)+((num+2))*value, or shorter (2*num+3)*value. Each time, there is a little discussion needed - if the previous joint sets a new origin point (parent/child) of if that is part of the chain or not. Here the hierarchy is half of the result, not in numbers but in “relative weight”. With all of that, I attach an file here, even knowing that one could argue one or the other way, as with the hierarchy (again) we introduce one more element in the equation—which will not add up in the degree values (given the parameter fields) But I think everyone familiar with Fibonacci and the C4D Hierarchy will see the problem instantly.
XPresso would solve that more accurately, but more work, you pick. Perhaps using the Fibonacci numbers in a list…etc. Long story, and tell me the need/actual use, and I will dive perhaps deeper into it. ;o)

Enjoy