melvelvin
patreon
WIP#21: More Better Men
🕑 Added 2022-08-13 20:44:02 +0000 UTCHey all! This week I've taken a bunch of steps to more permanently improve my male models again! Once again let me just apologise up front for not writing as many wip updates as I should... I'm always so busy doing stuff that I forget to document anything! If you'll recall in the previous update I was having a great deal of difficulty maneuvering big mac's big meat, and her you can see my solution to that problem! More on that later.
First, in finally making (two thirds of, at time of writing) a proper model for Soarin, I finally took the plunge on making male models that aren't extremely jank. Previously all my males were hacked together from other males, which themselves were derived from some probably daz genesis base model. No longer! Soarin is pure 100% Melvin topology, just like the girls, and that means he can benefit from all the bits and pieces I built for them, like semi-compatable skin maps, corrective driven keys, and things like that! Also, topologically, I kept my meshes as sparse as possible, shown pretty well in a comparison between soarin and big mac, who is still using a genesis-derived mesh (much like nextgen does)
So what's going on with that schmeat? glad you asked. A lot of rigs for lewd males just go with a normal FK chain, which is only really an appropriate rigging method for a peen when no action is happening. As soon as you introduce an anchoring constriction (Like a mouth or other opening) somewhere along the shaft you really need an IK system to make sure everything is lined up correctly. However, the location of that constricting anchor needs to slide up and down the shaft. That's a hard rigging problem, and not one that you normally need to deal with when rigging an arm or a leg - knees don't usually slide up and down their leg! Something a little trickier is required.
Previously I tried using a spline IK system - on the fluttershy bic mac deepthroat animation (coming soon!), and it had some serious downsides. I got the motion I needed, but the setup was cumbersome and worse, it was actually too much for blender's dependency graph to handle, meaning it wouldn't update correctly as the pose changed. I worked around it for that animation, but it can't be a permanent solution. So IK doesn't work, and splines don't work, so what did I do?
Well, I went back to basics on what IK actually is. At it's core, an Inverse Kinematic solver takes a destination position and then applies a series of geometric equations to a chain of two or more bones so that the end of that chain is as close to that destination position as possible. So, I built my own, with drivers and constraints! This is a simpler, more nice IK case where the bone chain always follows a path corresponding to a circular arc - that's how we get nicer, subtler bends in the member. I had to do a lot of trigonometry to make the expressions required! Then, since it's spread across multiple drivers, this "bootleg IK solver" plays much nicer with being partially overriden, so each bone calculates where on the circular path it is relative to the constriction, and if it's past that point it disregards its IK rotation, and instead copies the rotation of the target bone, so that the shaft is predictably straight once inside whoever is being penetrated. That, in combination with the shrinkwrap/doughnut/surface deform technique for handling holes deforming around penetrators (not my invention, I can't find the source at the moment I'm afraid), we're looking at a new era of stable and predictable interaction! Exciting!
