Hello! Made a first Demo;
Didnât know where to post, but here seemed to be fair enough:
Best Regards,
Pxy-Gnomes
5 Likes
Looks like a pretty good eye rig. 
Thank you;
Iâm still researching some plausible improvements.
In this particular version/iteration, I have decided to make that the Meshâs Rest Position (Mesh Object on Edit Mode) for the Eyelids, start with the Eyes Open, instead of Eyes Closed.
This makes it more difficult, especially on this Character Design with complex Eyelashes shapes, to have the Close the Eyes with easiness, without the help of something, like a very specific Pose Asset for example, or (something Iâm considering yet) a single, special Shape Key for Eyes Closed shapes (which would serve as a âfake, secondary Rest Positionâ, if you see what I mean).
On the previous versions/iterations, because the Eyes starting on an Eyes Closed (Mesh Rest Position) instead, naturally, the struggle was the opposite: I had to find the correct Deforms on an Eyes Open Pose; but this was much easier to solve even without Shape Key because there was more freedom âit wasnât the case of an âencounterâ of Meshes, notably Top Eyelid/Eyelash with Bottom Eyelid/Eyelash in the case of making a Closed Eyes Pose.
Even though this worked better, there was a counterpart that I hadnât been able to solve (but again, a Shape Key would have solved probably): The Eyelids Mesh, which is simply made of regular Face Rings around the Eye Holes, would not naturally Collapse as the Eyes Open; instead, they would âenterâ inside the Head. The viewer wouldnât realize this particular, unnatural event, unless looking at what was happening to the Wireframe for example. 
It didnât cause any bad Deforms, as it was just (Eyelids) Mesh going behind the curtains; but this wasnât satisfactory and annoyed me, because to solve a âgradual Top/Bottom Eyelid Mesh Decollapsing Deformâ, with Armature Deform only, would be (probably) quite complicated.
So, although everything else seem to work fine, there would be at least this fundamental challenge for this Eyes Rig: which is different challenging, depending of the fact if we start with and Eyes Closed Rest Position, or an Eyes Open Rest Position; they are unique, different challenges, but they relate to this same, Rigging major problematic: the choice of the Mesh Rest Position.
Actually, by just writing this essay, gave me an insight to explore. This makes your feedback even more valuable.
Perhaps, if I attempted a middleway Open/Closed Eyes Mesh Rest Position, it might be possible to obtain, simultaneously and to a certain amount, the Deforms advantages of both of the Mesh Rest Positions (Eyes 100% Close, Eyes 100% Open).
Making changes to a Rig like this one, unfortunately can take a lot of work and time; so this may be considered a disadvantage. But now Iâm willing to explore that mentioned proposition.
PS: Currently, making tests with Corrective Smooth Modifier for the Eyelashes Meshes; this, or additional Subdivisions, kind of improve the Meshes matching, on the Eyes Closed Pose.
1 Like
Agreed. I havenât always done this myself (or even seen it widely used) but I think maybe the ideal rest pose for all joints and areas of deformation would be the midway point of any range of motion.
That way youâd tend to get both a bit of stretching and a bit of compression in the mesh at the extremes of fully open/extended and fully closed/contracted poses or bones (as happens naturally), but not so much as trying to go from one extreme to the other.
Hi, @Cire
Iâm sorry for the delay.
Although it could be considered a Rigging fundamental, this is not a recurrent nor explored topic. Iâve just started making experiments in the matter. Apparently, popular Rigging culture just sticks with the T-Pose/A-Pose problematic, taking it for granted, and then we end up not even minding why we do those, nor if there canât be anything more adequate.
This is no guide unfortunately, but itâs just a proof of concept as I was intrigued by the possibilities.
I believe youâd like to make some tests if you have time, but try to keep things simple regarding Boneâs orientations (Iâll explain this later). So far, whatâve been testing, I can tell that proposition is good. But to find the âidealâ Rest Position for the Mesh, is where it starts to become more complex:
-
In terms of plain Deforms, it is true: the Rest Position of the Mesh (and not of the Bones) determines the Mesh starting Deform condition for the employed Skinning method(s) to the Model. Unfortunately, It doesnât seem possible to have more than 1 single Rest Position for the Mesh Object at a time, so this means that, if we make an extreme Rest Position, it would certainly give the âperfectâ Deform Pose on that starting Pose, but it would also cause more horrid Deforms on Poses on the other extremeties. Therefore, thinking of the exact middle Rest Position, although seems logically good, actually isnât organically good, because certain generated Artifacts are easier than others to be solved through Corrective methods. So this means, depending on the anatomy and the specifics of the Artifacts that are more or less troublesome, we might prefer not an exact 50% Rest Position for the Mesh, but perhaps 75% towards one of the extremes, or whatever arbitrary %. Certain anatomical shapes, we may rather have them already proehiment from start (Mesh Rest Position), like in my exemple, it could be Eyelids 80% Closed, instead of halfway 50%? Or anywhere between 30% to 70% Open); I think, it depends, one one hand, on the benefits the Rest Position may, immediately and holistically, offer to the Deforms, and on the other hand, on how this helps avoiding reducing the worst Artifacts we donât want to face (while perhaps having to face other Artifacts/problems, that migth be easier to fix).
-
Unfortunately, there are additional problems that can occur regularly, as we try to work with different Mesh Rest Positions. These often require specialized adaptations for the Sketeton; as we wouldnât have to deal with the adaptations, if we were dealing with more straightforward, simple Rest Positions, such as the conventional T-Pose for example.
Itâs the Control/Mechanism Rig, and how the Armature (Skeleton) relates to the Mesh. The Bones Alignments, Bones Rotations/Rolls, and it starts to get crazy if we start thinking about Bendy Bones Rolls and Limbs Twisting features, things can start to get quite complicated to solve. I was facing a big complication in an Arm Righ with these Skeleton problems (Bending Bones Splines with Limb Twisting features were too difficult to solve), that I decided, not to give up on a different Rest Position for the Arm, but to simplify the Rest Position I inteded for the Deform Rig, so to make it simpler for solving the Control/Mechanism Rig. It would be like a T-Pose for the Arms, except the Arms would be up like nearly 90Âş, and slightly bent, but making sure Upper-Arm+Forearm+BackOfTheHand, all Limbs had the same orientation (facing up), so I could start out with matching Alignment for both Arm Mesh and Arm Bones. I still havenât completed the attempt, but Iâm confident it has a value to simplify certain processes if we can. On other occasions, what I ended up conceiving (probably not a new thing), were âWarping Bonesâ; in principle, the purpose of those (usually Main/Master) Bones would be to automatically set a regular T-Pose on Pose Mode, from any Rest Position that we may have for Deform Rig. The problem of Warping is that, we may not be able to immediately track all the Transforms and Deforms that would be occuring, nor problematic relations between Transforms & Deforms. It also requires some thought for adapting Warping to certain Bone structures; like, for example: if we have an Advanced Foot Rig, and the Rest Position has the Feet Mesh in a very crazy Position and Orientation in World Space, should we have the Advanced Foot Rig Aligned with those crazy Positioning of the Foot Mesh? Or should we get it Aligned with World Space? Well, both solutions work, but they need proper adaptation (Aligning [most of] the Mechanisms to World Space like T-Pose would be better), but the Warping system must also be good.
I think maybe youâll understand now what I too more time to answer. 
But you are correct indeed: Rest Position can be very impactful; Iâm convinced, with more investigation, that it is possible to achieve great solutions that should be a match to what is thought more conventionally. Thanks for commenting on the subject! Take care.
1 Like
Fascinating! Great job
1 Like
I love this eye rig. Definitely will pay off during animation. Iâve never attempted anything that complex, so I have to waste time keyframing regular bones and trying to force them to behave
2 Likes
Thanks!
I have a (very, very) long tutorial âIntegrated Eyes Rigâ on my Youtube channel that shows the process of making this kind of Rig. (Better skip the âFundamentalsâ section, the first videos).
The tutorial would require a few tweaks, as over time I got some improvements (especially to the Control Rig), (and I found out I shouldnât Weight Paint the whole Head to the DEF Head Bone at 1.000 because that would lag the Mesh behind in relation to the Deform Bones, offsetting my Controllers⌠meaning I had to double the efforts to get the Pose I wanted; actually, this whole DEMO was with this kind of offsetting). Overall, Iâm looking forward to making a 2.0 with updates, but this wouldnât be a project for now.
But the most important things were already on the original, such as:
I use Shrinkwrap Modifiers, for example to avoid Interpenetrations between the Eyelids and the Eye Surfaces.š
The Mesh Topology of the elements of the Face, as well as the generic concept of how the Bendy Bones Splines Sequences arrangement and choices are made to match that Character Design, were relative accurate (but more recently Iâve made valid simplifications (which are present in the DEMO and not in the tutorail: I think I ended up Deleting 1 BB Spline from the Eyebrow Rig and a few from the Orbicularis Oculi Rig). In the tutorial, the concept of having just 2 BB Splines for each Eyelid (Top Eyelid and Bottom Eyelid) as separated BB Splines Sequences, prooved to be quite solid over and over again; while the concept of having a Looping BB Splines Sequence around the Eye Holes (the Orbicularis Oculi muscles Region), is usually acceptable, but I couldnât figure at that time how to make a good Control Rig for it (which can take some time to figure out, and there is a multitude of possibilities).
The Weight Painting techniques presented, with proper Paint Masking and Vertex Selection require a certain practice,
The Automated Eyelashes Rig is very specific; unfortunately, it takes a lot of back-and-forth to fix sometimes, but itâs reliable regarding quality. Iâm currently willing to find out new alternatives to it; I believe it has potential to be simplified.
Yes. I think those would be the most relevant to know technically.
š Shrinkwrap Bone Constraint can be used safely for similar results, so it does work as a substitute (itâs not as precise, althought it does open up doors for Baking and Exportation just in case that would matter). Spline Skinning, with Bendy Bones, can be substituted by a âlegionâ of Shrinkwrapping Deform Bones with Auto-Weights (plain Linear Blend Skinning method) which would move in synchrony in a sort of âProportional Posingâ way (usually Bone Constraints with varying Influences), that would require to be built.
1 Like