Soulspire Dev Update #6

Happy Tuesday everyone! I'm trying to do weekly dev logs, but I totally forgot about yesterday's because I was busy working on the first animation in the game - hope you all don't mind. 🙂

Version 2.5 is mostly done at this point and is on track for a release this Friday (March 20) for high-tier patrons and on Saturday (March 21) for all patrons. This update will be basically linear, covering all the events between the end of version 2 and the gameplay elements of version 3. I can't really say too much more without spoiling things, but version 3 will start featuring the reverse trainer gameplay, so you can imagine that there are some blanks to fill in. 🙂 It's still a pretty decently sized update for less than three weeks of work (5500+ words) and it'll feature the first animation in the game. I hope you like jiggling!

Speaking of jiggling - I know some people are curious about the process itself for things like jiggle animations, so I'm including a summary on what I've done in Blender so far to figure out this new animation stuff. If you're not interested, then all the game release stuff is above, so you can just skip this part. 🙂

Though Blender supports soft body simulations, I was actually recommended (by Puggy, who else) to use Blender's cloth simulations for jiggle sims. According to him, they're faster and still look great, as long as you set things up correctly. Obviously, you can't use "normal" cloth parameters to simulate jiggling body parts, so the simulation parameters are similar to a slightly elastic cloth bag filled with a dense fluid. I'm sure you can imagine how that would translate over to a jiggle sim. 🙂 Tweaking the parameters gives you more or less jiggle, basically, though the specific effects of each parameter (of which there are literally dozens) aren't clear to me at the moment.

The other important part of the simulation is the "cage", which is a low-poly version of what you actually want simulated. Once you create a cage and fit it to the jiggly parts, you hook it up to the model (so that it moves with the model) and run the simulation on the cage. The last (and easiest step) is actually transferring the simulated motion from the cage to the model. As I learned, the cage has to be constructed very carefully, because every aspect of the geometry of the cage (ie. how many faces it has, whether those faces are triangles vs. quadrilaterals, the angle of curvature, etc.) influences how it'll simulate. My first attempt at a cage was a complete disaster, but Puggy sent me a known-good cage that worked like a charm. Thank you Puggy!

With a properly constructed cage and good simulation parameters, all that was left was animating the model. Blender's animation tools are quite good (better than Daz, though that's not a high bar), so it didn't take long to get everything moving how I planned. Running the simulation took less than a minute for a few seconds of jiggle, and it turned out really well! As it is with most things, the hard and time-consuming part was the final 5-10% of refinement, where I was tweaking movements and sim parameters slightly to get everything looking exactly how I wanted it to look. Now that I'm happy with the results, I've got the animation running as I type this dev log. It's about 10 hours through the 20 hour run time I'll need for this animation, and so far it's looking how I imagined it! Hopefully you all will enjoy it, too. 🙂

If you found this behind-the-scenes interesting, please let me know and I'll include more of them in the future! If you hated it, I guess you should also let me know that. 😛