Update 34: No Half-Measures


Hello all!
I'm going to recount the events as best I can that we went through this month as development of the scene continued, there's quite a bit to talk about!

Week 1: The Audio


To begin with, after careful review, I found I wasn't able to make the original voice work I'd had done work with the scene.

This was disappointing, but I'd had a gut feeling on an initial review that it wasn't going to fit, I had no complaints about the standard of it, It's just one of those cases where my vision for the scene didn't match the VA's interpretation, this is just something that happens sometimes, after a couple panic attacks about it, the first order of business was re-ordering voice work.

Enter Sultry Lamp who, after providing a sample which,1 if I didn't know any better, was just clipped from the actual Kiriko intro cinematic, delivered a wonderfully sensual, erotic, detailed and engrossing performance which slotted right in with how I saw the scene playing out.

At the attachments at the bottom you will find a WIP Mp4 which has audio in it for an example of the work, it truly is a fantastic performance, and has been a pleasure to work to.

This was resolved quickly, but it cost me about a week before I could really get started, as the entire animation is based around the audio performance, the new audio had to be delivered, cut, and prepared for use, this cost us the week.

Week 2: The Environment


I turned my attention in the meanwhile to establishing the environment layout. I ended up with something that had a rough idea of the space and shape I was going with. But to call it basic would be being generous:
Several hours of fiddling with this, and a growing impatience to start animating was distilling my attention here, luckily my editor managed to step up and take on the job for me, He was able to continue work on that, building not only this environment, but a modular asset pack we can use on all productions moving forward:
I seriously can't overstate how much better this guy is at this stuff than I am
It is thanks to the support and patience of my supporters that I'm able to not only attract, but pay this kind of talent. I can delegate the task to him, pay him a fair and living wage to do so, whilst I turn my attention to the area I'm best at:


Weeks 3 & 4: The Animation

This is the first time I've returned to a female (genitalia based) participant since Mercy X Aimee, and I wanted it to be -good-. Mocap and a decently thought out character rig promised that possibility, and thus I got started

We start with a mocap take, in this case, I've captured and transferred (after some adjustment) the movement of the hips and shoulders onto the character:
Yeah you're gonna have to deal with the creepy dead eyes and no head movement for a little bit here, it gets comparatively worse before it gets better!

Next we add breathing, the mocap setup is not accurate enough, and it's not really practical to get this kind of movement in a mocap scenario, so this is always a manual process:
It's not straightforward either! Here I'm finding the beginning and end points of the breath, then adjusting the curve based on the aggressiveness of the motion, we tend to breathe in steps, IE you release the breath rapidly at first, and then slower as you expend the air volume, likewise in reverse, you tend to sharply intake the volume at first, then fill the remaining volume slowly.

Especially when you're being assaulted by variable, precision applied pleasure.

It took about 2 days at 10-12 hours a day to get the breathing pacing done for the entire production.
This starts to make the motion feel a lot better! But I felt there was more that was needed to tie it all together, thus I created a controller which I then used to add a variable state of tension to her body:
The idea is this key allows me to adjust globally a few values which make her more tensed/at ease, so it changes a few things, raises shoulders, brings elbows in, tightens thighs, angles the hips, pops her abs out.

Another couple of days and that layer is then added to the entire timeline:
At this point I can consider the body animation pretty much done, it is devoid of the proper physics based responses, but for the most part, this really ties the entire body motion together.
Thus it's time to animate the things that are causing her to writhe so:
After another couple anxiety attacks at the scale of this job, far more robust rigs then I'm used to working with made this task fairly trivial, albeit time consuming, as it's a very curated, artistically driven motion. The feeling I wanted for the arms was to commune precise, but organic movements.

Worth noting at this stage that this doesn't feature her skin reactions to the arms, which is something we're going to tackle at the finalisation phase once I'm done with all the key animation, more on this later.

Next up, as I'm sure all of you have noticed, no mask on her, this production has mouth animation, voice lines and such, thus, facial animation is needed!

Now this is the first time I've worked with facial capture, my initial use of it felt a little bit underwhelming honestly.
Looking from the outside in, it would be understandable to think that mocap is simply a case of strapping in, recording the session and slapping the result onto the model. As is probably pretty clear by now, it requires far more curation and adjustment to make something that looks good.

Head motion is an interesting one, this attribute is tracked by the suit, I even did another separate take of the entire scene just to capture it, but you never know what direction you should turn your head in at a certain point, because you can't really consider how that's going to look until you have your angles all figured out, thus animating this cannot be anything other than a manual job to look it's best.

Luckily thanks to my technical lead, I have an excellent rig that allows those kind of adjustments on top of the facial capture:
This is easily the most time consuming phase of the animation, but as is hopefully obvious, absolutely necessary to bring the entire motion together into a cohesive whole.
It's this stage where I embellish the little details, fix anything erroneous, and really just polish the animation. This includes animating the head, eye targets, and making any adjustments or corrections to the mocap and lip sync.

The facial mocap adds an excellent base, takes care of a lot of the timing, and once curated together with manual animation, really sends the facial animation to a whole new level, Strange isnt it, that on it's own it doesn't seem like that much!

And whereas this process is normally the most time consuming part (you can polish forever), it only took me a week to get halfway through polishing the entire 12 ish minute production to this standard, which is pretty incredible honestly, given this scene has a lot more facial animation in view at any one time than any previous production I've undertaken.

Previously it would take me an entire day to do maybe a 20 second shot to this standard. The facial segments of the previous Brig animation took me 2 weeks to complete, and that's not even half of what I've done on this piece already.

At this stage this shot is pretty much what I'd consider key animation complete, There is a physics pass that my technical director will be adding, which will essentially provide the skin reactions to the arms, along with jiggles in the right fatty places.

Some parts though are extra jiggly, so they benefit from a little live wiggle:
Boobtacular as it is, there remains pockets of fat that would really benefit from extra jiggle, and a skin simulation would also sort out areas that are not being tackled correctly currently

Once I'm done with the key animation for the entire production, this layer will be added, along with all the rest of the parts that need simulating, mainly the hair, and the chairs cushions reaction to her body/head movements.

Be sure to check out the attached mp4 which has sultry's -decadent- audio performance, and also plays in 24 fps as opposed to the embedded GIF's in the post.



The Conclusion

So I spent a lot of time on this update because, I was very aware of the fact that it's been a while since everyone's seen some new content from me.

I also slipped into the old habit of making estimates on timescales, whereas this is a skill I've proven to be very bad at in the past, I hope that shedding some light on the production process goes some way to explaining why the nature of these long format productions is so prone to unanticipated delays, and why estimating things is far more challenging then it might seem.

I lost a lot of time overall doing everything but the part I'm best equipped to tackle, this is something I am going to take steps to improve as we move forward.

In my endless pursuit of perfection this production represents a landmark moment for me in many ways; I am finally getting -close- to exactly what I see in my head when I imagine these scenes, the models I'm using are properly optimised and the environments are going to be better than ever. it is extremely exciting to see unfold and I am incredibly excited to share it with you all when it is ready.

As excellent a job as my technical director did preparing this model for me, it is still one based on limitations, as good as this model looks, it will pale in comparison to the standard we're working on, let us look to:

The Future


I can't honestly talk about this and accurately communicate the level of passion for this project my technical director has, so I asked him to write a few paragraphs about what he's working on right now:

One of the biggest hurdles of 3D animation when aiming for humans with realism is the uncanny valley that most rigging methods will produce!

Single use, linear shape keys serve as patchwork solutions and substitutions for properly placed joints whilst contributing to file size bloat, many models are built on spaghetti solutions built to tackle problems produced by a lack of anatomy knowledge.

Big movie studios resort to muscle simulations when making realistic creatures to help take away that uncanny valley by literally simulating how the human body works in real life.

One of the best examples I can give you why this workflow comes in useful is James Cameron's: Avatar

Only one Navi was made with the inner skeleton, muscles, veins, etc… From that one base they were able to generate all the different Navi characters!

Most people tend to believe that blender being free is incapable of such rigging but this couldn’t be further from the truth! Blender has everything needed to create such a rig.What we are currently working on here at MJP will hopefully help people finally believe it is totally achievable with the right approach and a shite load of research & development!

Like any muscle simulation, we need a good accurate skeleton to drive the entire setup! Since humans are something most of us are used to looking at in films & in real life, there is a lot less room for creativity than doing a dinosaur or creature for example!

I’ve downloaded and bought quite a few human skeleton assets just to be disappointed with how most of them are not very accurate unless we spend a fortune on them! Also not a good idea to go dig up a grave just to 3D scan one either hehehehe!

So in this particular case, the best possible course of action was to take the time to study every single bone as madly detailed as possible to recreate an accurate sculpture of the human skeleton, as close to medically accurate as possible with the available references without going completely nuts on details!

Since the skeleton is used to simulate skin sliding on bones, muscle anchors, muscle sliding on top of bones, tendons, veins, fat and such, I needed to make sure I knew exactly where those muscles and tendons needed to sit on the skeleton! So I painstakingly studied every bone to have as many of the major bone landmarks as possible so it can be as accurate as it can be!

So the idea is fairly simple (simpler than making it hehehe!): make an accurate skeleton model that is usable for animation, rig it with a fairly complex rig that allows you to change the shapes and thickness of the skeleton bones directly with the armature bones as opposed to using shape keys, allowing us to modify the underlying structure to fit whatever human character we need.

That also allows for better muscle placements as well as anchor points but also to allow the rig to be as flexible as possible while still giving us the ability to modify the overall look of the character but also to maintain complete animation control over motion capture!

Not going to lie, this is a LOT of work and sculpting this skeleton was very challenging but also very rewarding! Now doing the retopology and yes, it is just as challenging! We’re taking the time to do it the right way and we are very proud of its progress!

Jeannot (Jano) Landry
Aka: Blender Pirate 
Technical Director @ MJP


So when's it coming out?

When it's ready! More news when it's available, bye for now!