Thank you very much, smckenzie, for the reply.

My best wishes for your project!

Hi whereas-stuff,

Thank you for the file and information.

Hair collisions in C4D only happen while dynamics are running — the program steps through time in little slices (sub-steps) to check if hairs hit objects.
When you use the Brush while the scene is paused, you’re instantly moving hair from point A to point B with zero time passing — so the collision system never gets a chance to check.
Even during playback, if you move hair very fast, it can skip past an object between collision checks unless you increase sub-steps in the Hair or Project Dynamics settings.
For grooming, the Brush is meant for art direction, not physics — if you need collision during grooming, play the simulation slowly and brush gently.
Consider (later) enableHair> Dynamics> 'Rigid' mode to lock the style once it’s where you want.

In short, movement in no time is like infinite speed, like mph, km/h, or 1/0. In all cases, Simulations are not matching reality; they aim to find the best compromise between reality and computer power, typically with iterations or sub-steps, like 20 per frame.

All the the best

Thank you for your reply, resort-sand.

My best wishes for your project, and have a great Sunday as well!

You're very welcome, Hoi!

Thank you very much for the reply!

Enjoy your project

Happy New Year, Hoi! Thank you.

Thank you for the file.

Please have a look here if that fits your target:
CV4_2026_drs_26_MGrl_01.c4d

The idea is simple, but it needs some care.
The reverse animation must be in the keyframes of the Cube. It does something, and then nothing, then does the animation to bring it back.

In short, we have a three-act story.
Animating, nothing, then reverse.
Note that the reverse can be anything, not needed to do it exactly the same way, but inverted for the target position to loop.
CV4_2026_drs_26_MGrl_02.c4d

The animation Act One has a certain number of frames.
The Reverse (act three) has a certain number of frames.
How to define the "nothing happens" in the middle (act 2)?

The middle part starts after the first animation is done, plus the Step "delay", perhaps two frames (to reverse the Spline Interface), then the reverse part is animated.

The 40 frames in your example are always the max duration, independent of the Clone amount. Hence, the simplicity of the three-act story. If those three parts overlap, it will not work with the Spline Interface. Step Field> Remapping> Contour> Spline.

Is that working for you?

Cheers

Hi reading-card,

If the Connect is not in use, the weird jumping of the geometry stops.
The point numbers change constantly, and since the points are exactly on top of each other, which one is used is “survives the weld is random. The only way to make the points stable is to move the part that should have a stable polygon number a little bit, in this case, to negative Z, while also being the lower child of the Connect. However, that requires increasing the weld tolerance. Which in return has a lot of new weldong of points, and the change here is not wanted, of course.
A simple example of the changing Point ID numbers:
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_Clips_DRS/20260103_Connect_changes.mp4
Hence why I searched for something that works without Connect.

Perhaps check with tech support to get a second opinion.
https://www.maxon.net/en/support-center

Here is an idea to make the faces exchangeable in a short time. Just exchange it in the XPresso editor.

The idea is based on taking two loops from the "frame of the mask and using the B-Spline setting in the Splines.
To get the first loop of the face as data, it needs to be transparent.
With these two points per spline, a certain softness is given, but a small indicator for an edge remains.

I have typed in all values manually, as the mesh has no sequence ID numbers for the points.
While edges run in both directions and randomly
Also, here, the Connect changes the geometry along the animation.

Example (exploration)
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260103_CV4_2026_drs_26_CAfe_01.c4d.zip

The next exploration is an XPresso that takes the out Edge Point Numbers into a Condition again, to “normalize the point access numbers. Given that the face is a standard, the work to type them all in was worth it; however, that should be a feature request for the Selection-Iteration.
With a Polygon shape from a given Loft (N-Side with 56 “corners”) and using the “Current State to Object” to get it, the connection was done with an iteration of the Condition and the Face Mask.

Here is the example, and the face can be swapped out. Note that the Damping the in the Scene> Simulation>Simulation>Dampening is set to 25%.
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260103_CV4_2026_drs_26_CAfe_21.c4d.zip

The key idea would be to define a workflow that stores the PLA animation and applies it to the model as is. Meaning creating a model as needed, and applying the Morph Shapes to it. Then the PoseMorph would run it.

Good luck with the Project.

Thank you very much for the file.

The section that you indicated as a problem could be "supported" in several ways; three options are attached.

Forcefield, moving "Repel" (Attractor), and roughly aligned Gravity with a spline Field.

Examples
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260103_CV4_2026_drs_26_SIgy-rp-rb_01.zip

Would that work for you?

Cheers

Thank you very much for the reply, reading-card,

The problem with the Connect Object, and too many influences.
There are many forces pulling at the same time. The PLA defines the face, but partially the Cloth is contributing to it, and Welding might not always lead to 100% welding, which means the point amount changes. This is not working for the algorithm.
I searched for another way. I will not list the explorations of the past hours, but share what works.

My suggestion
Take the Face with the PLA animation.
Select the edge loop that marks the end of the “mask”. Extrude the edge by just 0.1cm, create an Edge Selection Tag, and extrude it again, and after deselecting the Edge Selection Tag, create a new one. Name both differently. (Perhaps later, some more. For now, I would like to focus on the workflow.)

Double-click on the first Edge Selection tag to activate its selection. Shift key down (hold) and click on the Point Selection. Then go to Main Menu Character> Create Cluster.
Repeat this for the second edge (Points, Cluster).

In the Cluster Object> Object> Offset, set the offset where the Edge should be in space.

Note that the S.Z in negative flattens the profile, while S.X and S.Y in positive scales it.

Create now from the original Points of the face mask a Vertex Weight Tag, name it wisely, this is for the Cloth.

Set up Cloth (to the Polygon object Face Mask), as you need it, and use the Vertex map as before, in the Mix Animation Tab.

Apply the Subdivision Surface on top if you like.

The setup is smaller, in my opinion, easier to handle, and can be adjusted with the Cluster parameter.

Example
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260102_CV4_2026_drs_26_CAec_01-02.zip

One point, set the Attribute>Scene> Simulation> Scale to the size the object has. 100cm is too larger for this.

Please let me know if that works for your project

Cheers

Thanks for the reply, resort-sand, and. the kind feedback!

They are all there, so far I can see, but perhaps too small? Please note that I set the Sampling (Render Settings> Redshift> Sampling for previews typically to 0.1, which is OK, but not production quality. Better is 0.01 or even 0.001, but the render time increases drastically.

Please investigate these suggestions, so we can find what is not working on your side. Please take one step at a time.

1. Move the slider of the material to the right (delete the animation)
2. Set up the Sphere size to 10cm, for example
3. Cache the Particles with the option on the Folder: Particle group

Any insights that lead to solving this problem?

Happy New Year

Hi reading-card,

Thanks for updating the file. Zip isn't working by default here, safety/IT-wise. Files in C4D format up to 1MB are direct upload-able.
Dropbox is perfect (so is WeTransfer, Google, Adobe, and Apple cloud services). Please upload uncompressed, if needed again. Thank you.

Here is your file back, not considered the final version, but to show the problems. The face border is moving, the Cloner has to prep the result (MG Baking did not improve the results), the Loft Generator builds the Geometry, which is then again built in the Connector, and again In the SDS, while PLA is moving the face, and the Cloth is then to move everything, except for the Vertex map data.

I have gotten to the point of no Mesh-Chaos.

Here is the file:
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260102_CV4_2026_drs_26_CAcl_01.c4d.zip

The question is, do you set up a Pose morph for the Face from the Moves by Maxon? Just to get the full picture.

My impression is that I need to find a different approach. Example:
https://projectfiles.maxon.net/Cineversity_Forum_Support/2026_PROJECTS_DRS/20260102_CV4_2026_drs_26_CAcl_11.c4d.zip
I extended the Face as needed first, harvested some expressions, to added them as Target to the PoseMorphs. The point mode in the PoseMorph needs an on and off switch, which erases all point data, but not the morph target animation (sliders), but then it works. The vibrate tag is only there to trigger some motion for Cloth.

I will have a look in the morning to it.

All the best