In Depth - Lexie Walking Animation (Part 2) (Patreon)

Been a bit, but here we are again. As I wrote in part 1, this part will be about breast physics, clothing animation, and hair physics. 

Be warned, there are a lot of .gifs and images embedded, which might not play nicely on mobile and/or slow connections.

Breast Physics

Through the years I have been able to make 'the goods jiggle' through various methods. That's probably my favorite part about these 3D modelling applications, you have an entire toolset with more features than you'll ever need. Almost every step in the process can be handled in a number of ways. And that also applies to breast physics! In the past I've relied solely on bone movement, simulated ball of cloth imitating a breast, and even a mix of the two.

However for Lexie here, I used a more professional method by using Ziva Dynamics. The plugin allows you to simulate bones and muscle moving underneath the skin, but also fat! For the program that's just without any muscle and bones, which technically is accurate. I first caught wind of this plugin existing through Augmero's Mercy animation, specifically the pushup one. They've also since then, made a post on the workflow which has helped me a TON in figuring out how to even use the plugin.

I'll be honest, I don't know if I can say that I really know how to use the plugin. Despite this plugin having many more steps than the previous methods, it feels a lot more consistent because it is actually simulating the entire body. My other methods basically just approximated it, which often produced results that didn't really match up to the shape or behavior.

In the past I have tried countless times to animate something, but any motivation often collapsed here. Because of the fact that the results were all over the place. To this day I have no idea what I was doing wrong, why the legs are freaking out so much. There always seemed to be going something wrong somewhere. Luckily for this animation, all I care about is the breasts.

You might think at this point, that they don't really move that much? Yeah I felt that too, no matter what kind of settings I tried. Even the .gif down below, which is 100% the version that I use in the animation, seem like they haven't really much going. Though the .gif doesn't make it seem any better with it skipping a lot of the frames in between.

Thankfully I have always intended to cover them up under a shirt, to hide the imperfections. For that I simply used Marvelous Designer, a program that I pretty much do 99% of my clothing in. And something I have also covered in the past (sadly a lot of the images don't work anymore). Before I go to cover clothing animation however, the second thing you might have noticed is that the shape of the breasts are quite different from in the animation.

While I do use the breasts as a base, the shape that 'adopts' it is much fuller. For Skin Wrap it doesn't have to be a 1:1 shape, you can make it bigger or smaller than the object and the modifier will try its best to approximate it. Which is also why the shirt is such a godsend for this. I cover Skin Wrap in the final chapter of this post. So with that brought together, I made an cached animation that I can use in Marvelous Designer. Without the weird glitches from the Ziva animation, but with the jiggles and the fuller shape!

Clothing Animation

This was a pretty straight-forward process, probably the easiest out of the bunch. Mainly because many of my scenes already use the timeline/animation. Sometimes cloth (especially ones that drape like jackets) need to be animated into the pose to make it look natural. In order to make that happen, I make a quick A to B animation and export that.

And in the case of Lexie here, that animation happens to be a fully fledged one. And the program handles it like a champ! After that, it was simply a matter of exporting the animated cloth back to 3ds Max, texture it, and that is the clothing part done.

Hair Physics

The method used here is pretty much identical to my other animations, hair bones that reacts to the movement of the body via springs (its called spring dynamics in 3ds Max).

Don't take this as a reference for how many you need to have for hair to move realistic. I ended up not even using each bone as a 'spring' because each added spring cascades into the next and causing some extremely janky reactions when too many are in play from 1 'parent chain'.

This by itself is not too crazy of an implementation, but what I hadn't considered before is an animation that involves walking forwards. For the springy bones I had just created, that meant a constant force was being applied from the direction Lexie was walking in. My assumption I'm going to make might be wrong, but I believe this is just an exaggerated version of every boob physics version in games out there. They don't account for the constant momentum (and rely way too much on rotational bounce but that is another whole can of worms).

So I had to find a way to remove part of the momentum while still retaining everything else... The bones themselves I couldn't adjust, since they are all moving dynamically. What I ended up going with, was to completely negate the forwards momentum by applying exactly the negative of it (so the hair stays in place, but still moves up and rotates). In the example below, both hair objects are parented to the green box. But the hair in the middle also has another animation layer applied (which I covered in part 1), which makes it also follow the green cross. And that cross moves in the exact opposite direction (only Y axis).

The final part is to take the middle hair (that now functions perfectly) and have it still be attached the head. That part was fortunately simple enough by using it as a morph target. So the final hair object that is on the head, have it morph to the hair object that is fixed in place, and that's it!

Conclusion

In the end, I hope I was able to give a bit more insight in went into the Lexie animation. I generally don't go too in-depth and kinda explain the solution I used, which hopefully still helps anyone who might find themselves in a similar situation.

Unfortunately it is quite the time commitment compared to a standard scene, and a lot of time investment before you even get the chance at seeing the results, whether they are good or bad. Of course it's also a matter of finding the correct workflow, which I also haven't landed on just yet.