The plan was: with all production assets ready, do some minor rigging and then animate all the scenes swiftly (about 1 scene/day). The plan was wrong. There were a few things in the rigs of the main characters that I wanted to fix. You know small things, that should be done quickly. Not only did I end up completely recreating their rigs, after I started animating them, and put the rigs through their paces I found many flaws and had to go back to tweaking the rig – then after tweaking them I had to start recreating the scene with the improved rig, and so I was basically doing the same work over and over – this eventually took me 4 days.

Note: In this entry I’m talking a lot about how I rig my characters in Moho the animation software I’m using. If you want to try it out yourself, you can get the (less advanced) Debut version for 60 bucks, which is well worth the money spent for a tool that can get you very far already. However I’m using Moho Pro (13.5), so YMMV (no, I’m not getting any money for this, I just love the software).

Mowld’s Moho Rig

What I Learned

Mowld’s rig had some issues I wanted to fix. For one, I wanted to improve his head-turn to a more side view angle. This unfortunately took me forever and I’m still not happy with the result – the reason? I have never ever visualized his head from the side. To this day I don’t know how his head would actually look from the side and I think it’s one of these impossible things. Some characters just aren’t conceptualized as 3D figures, so there’s no correct way to show them from all sides, because some of the angles in their faces are simply impossible in 3D – that’s the freedom that cartoons and comics give you, but you’ll curse once you get to animate those characters. And curse I did.

Do not try to keep your drawing simple, by using as few points as possible. This ain’t no 3D low poly exercise and less points will not make the rig perform better, on the contrary, you’ll have to make more excessive use of Bezier curves with more extreme angles and curvature, which will come to bite you in the ass big time. Better add more points to get just the shape you want, try not to tweak the Bezier curves at all and you’ll be surprised how much more “well behaved” your vectors and curvature will be when you get to animate them.

Using layer reordering animation in smart actions (i.e. smart actions are movements you can setup and reuse, like a body turn or a head turn), will lock the layer ordering in your timeline when you animate. So if you ever need to manually adjust the layer order in your animation (“oh, I need the hand to be behind the torso, not on top!”) while using a pre-made animation like this, you’re screwed.

What I Fixed

• I did fix the head turn, but it doesn’t look great. I’m still not going to spend much more time on this, gotta make an episode, folks.
• I added four bones to his jacket, so now it can react to “gravity” (to a certain degree) and also flow in the wind (new Moho 13.5 feature!) – as used in the very first scene of the episode. However, I think the new wind feature doesn’t really cut it yet, because it just won’t go beyond a strong breeze effect, whereas I want the tornado, blizzard, black-hole-sucking-you-in kind of wind.
• I fixed a bunch of mouth shapes so that the interpolation works for all shapes and doesn’t “jump” (in order for interpolation of mouth shapes to work, all shapes in all layers used for the mouth have to have the same number of points)
• I created completely new “angry/shouting” mouth shapes that are a bit more exaggerated and work well with interpolation.
• Slightly reorganized the layering of shapes and the body layers, and I added points here and there.
• I fixed body turns, I left in the layer reordering for arms and hands, but never move them above the head. If I then ever needed the arms or hands to go over the head, I can just create a reference layer of those limbs (similar to an “alias” or “instance” – like a copy that always changes with the original) on the fly and put them wherever I want in the hierarchy. Then I’d have a second set of arms and hands, and I can simply switch on the top set whenever I need those arms to be over everything else.

ALIS’ Moho Rig

What I Learned

ALIS rig is simple, since he has no facial controls or hand switches. I wanted to improve his extremities though to get a better rubber-band look. There seems to be no perfect way, however, to make rubber-hose arms with Moho. The arc IK solver never worked remotely correctly for me, so I tried just using normal joints with 2 limbs, but give the vectors huge angles/curves so that they would bend in a more circular shape. Unfortunately extreme Bezier handle sizes always cause more trouble than it’s worth when you start deforming/animating them (as mentioned above).

What I Fixed

• I re-rigged ALIS almost entirely and I’m not entirely happy yet
• I compromised on making his arms really thick lines which will bend nicely in a rubber-hose way, but then I can’t adjust the bending angles of the shoulder edges for perspective effects. It’s a workable solution, but will look odd in some cases. Anyway, I haven’t found a solution I’m happy with yet… (By the way in the first two scenes I’m still using an older version of ALIS, check out the final episode once it’s done whether you can spot the difference).

Phungus’ Moho Rig

What I learned

I found that you should never resize layers inside your rig, but instead always change the size of the vector drawings instead. In the original rig of Phungus I adjusted the eye size by resizing the layer and I always found the eyes to appear to have slightly thicker or darker lines than the rest of his body. Which in fact was due to the “scale compensation”, since the eye layer was larger than 100% and the line thickness was numerically set to the same value as the rest of the rig, so it effectively ended up being thicker than the rest.

What I fixed

• I resized his eye layers to 100%, and rescaled the eyes using the vector transform tool, problem solved
• Fixed a lot of mouth shapes so that layer interpolation works, I also gave him a lower lip line. I hope that doesn’t change his image too much, but I think this way it’s more expressive
  • If you watched the pilot episode very attentively you might have noticed that the angry mouth shapes of Phungus never had layer interpolation enabled, and so the animation of his mouth when he’s angry is always not smooth, but more choppy
• I tweaked the dynamic bone settings for his stalked eyes to be less “stiff”.
• Completely reworked the head and body turn
• Refined the squash and stretch of his head to go from slightly squashing / stretching it to do the full stalked eye retraction snail thingy. It works really well and now I don’t have to tweak all the bones I can just hit a lever to have him become a frightened snail.

Gluk’s Moho Rig

What I learned

Well, he’s an entirely new character and I built his rig based on the Bloop Animation Moho Tutorial, which is excellent, but you need to have some experience with Moho to really understand the rig. A lot of the screen capture is sped up and unless you’re familiar with Moho you might not catch all of the details.

What I did differently

• Head turns / body turns only through vector animation, not layer manipulation. In the Bloop Animation tutorial they do all smart actions, head turns, body turns animating the actual points of the vector shapes. I used to do head turns mostly just moving, squishing or stretching the entire layer of individual elements (the eye, or eyelid, or nose), because I thought messing around with vector points could cause a lot of conflicts if you overlap different kinds of motion (like head left/right turn combined with head up/down tilting). I think both approaches are valid, but the Bloop Animation way – while being more work, you need to tweak the hell out of it – gives you superior results. You will notice that Gluk’s head movements are more fluid, seem more organic, less pre-defined (but there still are a few angles that look quite odd to be honest).
• This character also has a smart action for his hat. One for it flying up and stretching, and one for the hat to land on his head and being squashed. I just thought it would be a fun touch to have him always almost lose his hard hat (if it sits that loosely, you wonder what the point is of wearing it).
• The tutorial also gave me a better rig for eye movement & head turns. On first sight they might look the same, because all you see is a pin bone controller, but here’s the difference.

Phungus’ Eye Rig

This is fairly straight forward. You can simply set up two bones to control the horizontal and vertical movement of the iris, but when animating it’s cumbersome having to mess around with two controllers so instead you add a third pin bone (the circle in the middle). Then you tell the two bones that control vertical and horizontal movement to target the bin bone. Now you can hide the two bones and move only the pin bone to control where your characters eyes look.

• Pros: Easy to setup
• Cons: When you move the pin bone too close to the root of the other bones their movement can become jumpy and hard to control

Gluk’s Eye Rig

This adds another layer, but the principle is the same. Two bones for controlling the horizontal/vertical movement (the long ones) of the iris. Each bone gets a pin bone that will be the target. Now each pin bone will be setup to follow only either x or the y axis of the central controller pin bone. Once this is setup you’d hide all the bones except for the pin bone controlling the whole bunch at the center.

• Pros: proportional control of eye/head movement without jumpy behavior even when you get close to the root of the bones.
• Cons: you will have to use the bone translation tool (T) for this rig, not the bone manipulation tool (Z), which is a bummer, because most other animation you’ll be using the latter.

Narration Changes

In episode 1 I did use a voice changer to slightly raise the pitch of Phungus and slightly lower the pitch for Mowld. I just didn’t think it would sound distinct enough otherwise. During recording I usually have all audio effects on so I can hear myself in the final sound over the headphones. However, I can’t use the voice changer that way, because that causes a noticeable audio lag, and it’s disorienting to hear your own voice with 1/3 or 1/2 second delay.

Oddly enough after recording episode 2, I totally forgot to switch the voice changer back on, and it never bothered me much. So the final voices might end up being all just me acting, no digital magic-sauce added skullduggery. Maybe I’m just getting into the characters more, or it’s the practice I had, but it’s fine. The voices are fine and sound distinct enough. I always admired how Seth McFarlane can switch voices for Family Guy on-the-fly and play out full dialogues all alone. I hope can reach that level one day.