Cerbera

I haven't done this myself so far, but I suspect the best way to go here might involve cycloid splines, blend mode cloners and Field Forces.
Searching around those terms I found this tutorial from Insydium.
 
 
Obviously, they are using their own particle system, but there is no reason to think the native one couldn't also do it... though I am probably not the best person to advise on the specifics of that; I don't have much experience in Cinema particles yet !
 
CBR

There are, if we look carefully, a number of things subtly wrong with the settings in this scene.
 
As HP predicts above, Render Perfect on the sphere is one of them, and fixes the effect of collision deformer not showing in render if we turn that off.
 
Next we had some suspect settings in the sweep object (parallel movement should be off, Banking should be ON, and End Rotation should be 0 degrees), which fixes the twist in your path spline over the indent, and restores it to correct, contiguous circular form all the way along its length...
 

 
The mode of the Collision Deformer should be Outside (volume) and the capsule should ideally not be a child of it, though it doesn't particularly matter in this scene.
 
Likewise object order in the OM could be better bearing in mind that Cinema scans downwards through it from the top.
Below is more ideal I would say, and general good practice.
 
 

 
Next, the curvature in your path spline within the sweep isn't quite matching the curve of the indent in the sphere, resulting in some unevenness in that area. This can be fixed by getting a new path spline for the sweep, which we get by doing Current State to Object on the main sphere (once collided) and then Edge to spline on the centre edge loop of that, but first we have to fix the collision object, which should also ideally be a sphere so that we can choose Hexa type and thereby avoid the complex pole on the end of the capsule it replaces, which was previously confusing the collision deformer, and producing some wanky / uneven collision vertices in the main sphere at the apex of the collision point, which would obviously effect any spline you subsequently derived from it. In my version I wanted a better, higher resolution sphere, but not one that caused the collision deformer any extra work, so changed the type of that sphere to Hexa as well, which then necessitated addition of a Spherify deformer before the Collision Deformer (because hexaspheres are not mathematically spherical out of the gate).
 
And then lastly that went under SDS to give us improved resolution, and my Current State to Object was performed on the SDS instead of the sphere itself, for maximum spline matching.
 
Anyway, I have fixed all that in the scene attached below...
 

 
groove-on-sphere CBR Fix.c4d
 
CBR
 

No, I don't think there is anything we can do about that with a single setting, presumably because clone color is an attribute of the cloner object, so (correctly IMO) considered important to reflect in the viewport. It has always been this way as far as I recall, at least back to R11. That is not to say it doesn't annoy me sometimes, and I do wonder sometimes if we shouldn't ask for options in the basic / display tab that allow this not to happen, or at least not to preview in vp. But of course then we get into the massively complex question of what should happen instead, which is a nightmare if you think about all the possible options it could offer, and I have renewed sympathy for why it is left as it is !
 
The workaround(s) are simple enough though - we could just use a cylindrical field and a plain effector in color mode to turn one clone the highlight colour thusly...
 

 
or any of the other ways the cloner offers to do that...
 
CBR
 

I have checked this again, this time with the exact shape shown in the original post, and I can reproduce the error now. It is so far unclear whether the cause is linked to ngons, viewing angle or the Tweak Mode of the Poly Pen (changing that to points seemed to significantly lessen the chances of it entering new poly mode for example), or all of the above ! But I do have enough evidence to report it as a potential bug now, so will do...
 
Thank you for raising the issue.
 
CBR

I can confirm that until very recently there was a bug where having isoline editing on but the SDS cage hidden in filters menu was not working, and indeed that seems to be what is going on with yours - we can simply see both the isolines and the base cage. As you have spotted, it will be resolvable simply by disabling isoline editing. And I can also reassure you there is nothing wrong with your actual mesh, other than a few easily solvable tris that probably don't matter if left !
 
However, if I was doing that Sub-D I would go for much lower resolution in the base mesh, and all the quads, like so...
 

 
We have enough topo here to handle a fairly decent level of bending at the base mesh level, but if not, we can run the bend deformer at the same level as the SDS and bend that instead.
 
CBR

Several options here, but you were on the right track...
 
1. Displacer should be fine for this, but you probably want to use a circular gradient as the driving texture...
 

 
...in which you control the characteristics of the curve by affecting the knots in the gradient, and that white dot between them...
Also check the various modes of that gradient - Linear, for example might be more suitable for this than an exponential mode.
 

 
2. Bezier Primitive. Control the peak on a high res patch with a much lower res grid. You can think of this as a plane with an FFD pre-applied !
 

 
3. FFD On plane - the 'manual' version of the above.
 
4. Soft Selection. If you would prefer to edit a high res plane directly, then soft selection would be the goto here, allowing you to specify various profiles of diffusing components selections so that they affect adjacent (unselected) ones too. Dome Mode shown below...
 

 
5. Saving possibly the best until last, I would personally use Fields for this, implemented via a Plain effector which is a child of a plane, and has its deformation mode set to Points. This, I suspect is going to be most useful because if you use the contour controls in the remapping tab of a spherical field, it lets you control the character of the curve directly via a spline... like this.
 

 
...which would give us this sort of result...
 

 
So that's most of the non-mathematical approaches with the actual character / interpolation of the ramp defined in a number of different ways...
 
The maths-based stuff I will leave to someone more into that than I am !
 
As for the Atom Array aspect of your question - That will work as a parent of any of the setups above, but if you'd rather not wrangle the unnecessary vertex spheres in that, so will cloning cylinders to edges (scale edge 100%) or stealing the topology using Edge to spline (still available parametrically via a correction deformer) and sweeping the result.
 

 
CBR
 
 

Here's another alternative, using most of the methods I mentioned above, which remains quite simple, probably isn't the final answer in an of itself, but it does go to show that we can project things at a flat plane, generate a spline from the result of that, then place that spline under the original plane, which we can then deform however we like (spherical mode wrap deformer shown below) and then sweep and animate end position to get the shape being drawn as any deformations happen to the plane, which will happen properly following the deforming curvature...
 

 
We can also do the same thing and steal the segment lines from the plane as well to give us a deformable grid with even line thickness.
 
Obviously has some limitations, in that we do have to nail down the projection shapes and get their splines, but could be useful anyway because it is relatively easy after that step to animate spline based things and can keyframe both deformers and end point in the triangle sweep to sync path drawing and shape morphing...
 

 
As you can hopefully see, if you deactivate the wrap and enable the couple of bend deformers also in the group you can get your saddle business going on.
 
Project triangle CBR 01.c4d
 
I imagine Pose Morph opens up some additional possibilities here too, but didn't have time to try that. 
 
CBR

Volume builder is just another set of tools in the toolbox, but particularly suited to a number of specific situations. Modelling is only a part of what it does.
Pre R2025.1, it found a lot of use in making up for the deficits we had in the boolean department, but that aspect has become less important since the new Boolean Tool arrived, which in and of itself solves a lot of problems and annoyances the old one had.
 
However VB remains useful, thusly...
 
1. Fast iteration and concepting; it can be preferable to use a VB setup to quickly put together basic forms which might need changing or updating at any time before a concept is finalised enough to become a polygonal model.
2. Transitional Forms and advantages over Booleans; For all those situations where boolean operations don't produce rounded or beveled transitions between operands VB can do that very easily and adjustably whilst not losing any parametricity. Uniquely they do this without needing to base the result on topology like  bevel deformer on a boole would for example, and are therefore free from the artefacts that usually results in.
3. Combining with ReMesher to save modelling time; It is sometimes a lot easier to get a tricky base form started using the VB before remeshing it early on, and carrying on with poly modelling from there.
4. Infinite resolution; VB setups can be generated at any level of detail in a way that no other method can (OK, sculpting can too, I concede :)
5. Incorporation of splines, points, generators and matrices into modelling; No other systems within Cinema allow this as fully / directly as VB / VM.
6. Allows Noise-based and Field-based modelling; useful for producing the sort of meshes that are nigh on impossible to create any other way, such as a network of holes and tunnels in an Aero bar for example, or the internals of an igneous asteroid...
7. Unique Layer and folder-based hierarchy; with similarly unique elements like erode / dilate and multiple levels of smoothing.
8. Clouds / volume based elements; SDF is not the only mode here, we also have a fog mode which is how we can generate our own cloud and volumetric forms and the like...
9. Vector Mode; useful for spatially-based vector and field operations that can control and apply to forces.
 
And there will be more I haven't mentioned - the Volume builder in Cinema is a lot wider in scope than first appears obvious, and with every improvement and iteration it gets it becomes more useful (see cache layers / transitions etc), and its less desirable behaviours get further minimised and improved upon. As primarily a trad modeller myself I don't use it as much as others do, but there are definitely times when it helps me and I call on it when other methods fall short, or would take an unreasonably long amount of time when time is an issue and deadlines are looming...
 
CBR

The symmetrize command works BY duplicating your target side to the opposite one, so anything that is on that side is going to get mirrored. And that seems to be exactly what happens in your file. SO it does follow the Blender workflow to an extent, whereby it sets things up so you are ready to make further changes to one side, but using the Symmetry MODE rather than the object. If you then want to use the Symmetry OBJECT instead of the mode then you still have to manually delete one side to provide the Symmetry object with what it wants to work correctly. So not so much user-error as having incorrect expectations set by the way another software works ?
 
CBR

Ok, I figured out the 'why' ! If there is an unconnected ON port in the Xpresso node for that object it breaks the 'enabled' port for some reason (I imagine that might be a bug so will report it for you). But at least that is a mystery solved ! 🙂
 
CBR

You shouldn't have to grab that specifically - clicking anywhere in free space with the move tool should move it freely, right ?
 
CBR
 
 

Only my guesses, but I think that could be because that feature is primarily meant to be used with spline sketch to project guide splies directly onto the model where they are probably most helpful, which isn't possible with parametric splines (or at least not without deformer help etc). But may also be something to do with RM needing actual points to work with rather than a real-time mathematically generated procedural object ?
 
What are we like hey ? 2 am on Xmas day, and both of us are here ! All my lot have gone to bed, but night owls don't sleep much I guess, or at least not now ! 🙂
 
Season's best to you HP...
 
CBR

Only my guesses, but I think that could be because that feature is primarily meant to be used with spline sketch to project guide splies directly onto the model where they are probably most helpful, which isn't possible with parametric splines (or at least not without deformer help etc). But may also be something to do with RM needing actual points to work with rather than a real-time mathematically generated procedural object ?
 
What are we like hey ? 2 am on Xmas day, and both of us are here ! All my lot have gone to bed, but night owls don't sleep much I guess, or at least not now ! 🙂
 
Season's best to you HP...
 
CBR

In your version you need to be in the Sculpt layout, and then the bake function is here...
 

 
Cool model btw - sort of like a cross between a Thestral and what you might see a White Walker riding !
 
CBR
 

Welcome to the Core and Happy Christmas ! 
 
I fully sympathize - I can never fully 'get' the behaviour of those clamp constraints, and they never seem to do exactly what I want, so I am the wrong person to advise on use of those.  But I can recommend a different way, depending on what you need.
 
One very simple way to get a straight spline between 2 things that move is to position 2 'target' Nulls at the connection points of / as children of your main objects, then use the Tracer to 'connect all nodes' which will constantly draw a straight spline between them.
 
If you want to try simulation then Rope Belt is probably the way to go, but I would tend to avoid that approach, as it is quite difficult to art direct 'tension' using that and things have a tendency to be very 'floppy'.
 
CBR

I would guess you are changing it in the wrong place, and not here, which is the place you need to fix it...
 

 
...and indeed, you had that set to default of -981 cm.
 
CBR
 

The only option we have inside Extrude / caps is the Bevel Outside option but I don't think that does quite what you are looking for either. So in this case, I would say, yes; use the bevel deformer instead.
 
CBR

Righty - here's how I'd go about modelling that base...
 
We have 12 way rotational symmetry, and within that 2 way planar symmetry, meaning we only have to design topology for a nice small manageable section. And we don't need to go to the centre with that, because that is on a dual axis curve so we don't want any poles there. We'll make a patch for that later.
 
1. I started by outlining the main nubbin with as few polys as I thought I could get away with (whilst still keeping them mainly even)...
 

 
I am using HB retopo for this, but if you find yourself without that, then polygon snap will do just fine in a pinch, and you could even shrink wrap onto the CAD model to more rigidly stick to that surface as you work)...
 
I am working with symmetry here, so have only modelled the right side... you can optionally use the symmetry Hub / mode, but for technical reasons, I didn't on this occasion !
 
2. Next, let's patch that any old how, expecting to need to revise it later, and expand the upper border... mostly using poly pen here, and frequently utilising things like smooth edges and equal spacing to keep things neat...
 

 
3. We need to continue out to half way across the square block bits, so did that, trying to keep edges as even as possible.
 

 
4. But now my modelling spidey senses kick in to suggest we'll be needing some harder edges and corners on the inner edge of the square bits, and that we should continue the lower loop around the indent on, to keep its outer control loop flowing around it (rather than down towards the centre), which led to the following topology changes...
 

 
So I've run an extra loop across the square bit, and slid some edges about to get tight control loops for those corners, but, crucially, have got 4 evenly spaced edges at both inner and outer borders, which is vital for perfect SDS circularity (top) and partial sphericity (inner). I actually went on a bit beyond this to tighten those corners further before the next stage, and added some consistency as to how I did the cornering and additional loops round bevels for better base mesh definition.
 

 
5.  So, having got one segment of the 12 mostly right, we can radial clone that x 12 to get the complete outer ring for the base. That'll  need to be in a connect (weld on) to join those together. Having done that, we can make that connect editable, and count our border edges, which in my case was 48, which we can divide by 4 to conclude that we need a 12 x 12 plane for that centre section, and we will need to make that editable, Fit Circle it, relax / iron the inner points, project it to the CAD model (shrink wrap deformer) so we match the curve exactly and then CO&D that to the connect we made earlier, which should match perfectly if plans have gone well thus far...
 

 
This means we don't have any complex poles in the base on that subtle curved area at the centre, which is important, because even though the slope is quite gentle there, it curves in 2 directions at once, so triangles or pole counts more than 4 will produce SDS artefacts that will be variously noticeable and not fixed by adding more SDS levels. SDS L3 is perfectly sufficient for this unless we choose to use SDS edge weighting (not necessary here), which would necessitate an extra level of it.
 

 
...and that leaves us a very manageable 48 radial polys with which to work our way up the rest of the bottle...
We conclude that section with an Optimize, and checking it with Mesh Checker to hopefully reveal all quads and no borders, isolated points, or other errors ! 🙂
 

 
Hope that helps
 
CBR

I have never used the plugin so can't compare what we have to it, but this is the recently added Normal Editor we now have within Cinema (2024+).

 
Not sure if that can do what you need, but that's what it looks like, and is accessible from the mesh menu or via commander.
 
CBR

Ok, if you are going to move stuff then I think soft selection (a tab within PSR tools) is most useful to you, in conjunction with a few of our selection tools to make sure we are only moving certain things...
 
For example, in poly mode, you could start with U,L (loop select) to grab a loop high up on the arm, then do U,F (fill selection) to grab every poly under that to the end of the fingers in one easy click... and then you can put the modelling axis in the most helpful place for the operation at hand (axis mode or sliders under tool attributes), turn on soft selection in any component mode, set a suitable radius and start massaging things in the right direction, using perhaps the standard transform tools in conjunction with Brush / Smear tools to make things go where you want them. Here's an example of how you might do that...
 
So here I have a low poly model of a larger gentleman with his arms in wrong place. I have done the above and selected all the polys in right arm, engaged soft selection (surface mode so we don't grab his body too by mistake) and moved and rotated the axis so it lines up with his body in a way that will be most helpful to me then rotating it from that point, straighter.
 

 
So using mainly rotate and move tools I can now get his arm vaguely where I want it, like so...
 

 
and I can keep adjusting until it is mostly right, and then turn attention to the now mangled loops in the transition area.
For these we can move into Edge mode, get the Move tool, and double click to get individual rings of edges that we can then move about to even out the transition and hopefully avoid any intersection.
 
Whilst we are in the modelling tools though, a number of other things there are likely to be very helpful; Iron (applied to the polys around the transition point) and Brush (smooth / relax modes) both do superb jobs of smoothing and untangling any bunched up or intersecting polys. That way I can get his pose exactly right, more like this one...
 

 
The important thing whilst doing all this is not to actually add or delete any points / edges or polys to this mesh, which will break your UVs if you do - the mesh's point count and order must remain the same in order to preserve that, but apart from that, you can do pretty much anything you like to your base mesh before you rig it !
 
Hope that helps
 
CBR
 
 
 

That'll be a spline interpolation thing. Change Natural to Uniform and increase Number to something like...

 
You could use less subdivisions if you only had more original points in the spline, but you don't so we have to keep the number high to stop it rounding off the corner. But now everything should be going along it at a consistent rate.
 
CBR

Alas my job went on all night, so only had time for a quick preliminary look at this in which I tried the plan above, applied a single level of SD and then changed the edge flow around the mouth. Not sure it's the holotype solution here, but hopefully it's a nudge in the right direction..
 

 
I may have an improved version of that later... but actually the more pronounced corner to the mouth is more what snake mouths do than what you were aiming for.
 
But we could also try and recue what you have and get what you want, but we really do need to add that extra loop to do it (and another one through the eyes nose ideally). But that is not something to fear, and once we have spread out the ones either side it's already looking a bit more even front to back if not top to bottom... and we don't get much pinching this way...
 

 
 
CBR

There has been an issue recently with selection in orthographic viewports, but I was not aware of anything affecting Perspective view. However testing that in latest versions today I concur with you there is a problem in all viewports, which I have reported. As you note, the only workaround I can find at the moment is to shift the view and reselect, which I appreciate isn't ideal.
 
If it is helpful at all while they work on that, in the meantime I believe it all works correctly back in 2023.1.2.
 
CBR

It's not my area of specialisation, but I watch developments with interest. I think most people still consider Maya the ultimate for rigging, but I don't think Cinema is doing too badly in that department either. We recently improved quite a lot about weights and symmetry workflows, and before that Character Object and C-Motion were good leaps forward in setting up these things quite quickly without a great deal of in-depth knowledge...
 
I found Noseman's recent 4 hour rigging series on the Maxon training team's YT channel particularly helpful in understanding all the rigging basics and possibilities open to us, so maybe it would help you catch up too...
 
 
Or for even more detail, and a better focus on character rather than than machine riggings, there was this excellent (and exhaustively thorough) series that preceded it.
 
 
CBR

I think the more apposite question might be 'Why is OBJ so much more inefficient than every other format?' !!
It's true OBJ files can be huge. I don't know specifically why this is, but feel sure it must have something to do with 'legacy', given that OBJ is a very old and 'first gen' format that should have been retired a while back, but (presumably) for backwards compatibility and 'library' reasons, hasn't been.
 
CBR