lordaardvarksfm
patreon
Progress Report - December 27, 2022
🕑 Added 2022-12-28 08:52:20 +0000 UTCHello, everyone!
I hope you've all been having a wonderful holiday season, as we quickly approach the end of 2020 v3 and quickly approach 2020 v4, the continuation of the hell year that never ends!
This week has been notably more focused and productive than the last few - and even somewhat on-topic, no less!
Really, my attention has been split between two things.
First up, the poster for this Progress Report:
Primordium!
For years now, I've been trying to come up with a Witcher project. I absolutely adore the main three ladies, and have wanted to do something with them for a long time. The problem has always been coming up with a journey for them with an appropriately bombastic end.
This past week though, I finally came up with a solution that I think is absolutely wonderful. And it's a pretty simple one too. I've been so focused on coming up with a journey with a respectable destination, that I never thought to consider having the journey be the destination!
And that is exactly what Primordium is - a series of short videos, each about 8 minutes long, where the journey itself is the focus. The ladies' objective is simple: discover the source of Primordium, an extremely powerful magical substance. But them actually finding the source isn't the point of the series - the hijinks they get up to along the way is!
I already have several episodes written out to varying degrees, with two complete storyboards, a partial storyboard, and enough material for roughly 10 episodes figured out. The first episode has actually already entered the voice-recording stage, after just one single stream seeing us put together the entire four-minute sex sequence for it.
And earlier today, I was able to lock in and confirm the full voice cast for the series!
InsideIncognita as Yennefer, Milly Stern as Triss, and Bordeaux Black as Ciri. Episode 1 also features Ivan E Recshun as a partner character.
These videos are intentionally designed to be relatively quick to produce, and I am expecting to be ready to install audio by the 7th of January. If things go well, it could be as early as the 14th that the video is fully animated, voiced, lip-synced, and ready to be sent off for sound design. Which means it could be as early as February that the first episode is ready to be released.
Now, the astute among you might recognize that date, January 14th, as being rather unique. And that is because that is the tentative date that my sound designer and I have figured out for the Claire/Jill video.
That leads cleanly into my second primary focus of this past week:
Auto-Jitter!
This one is a bit more abstract to explain to someone who isn't intimately familiar with my animating process, so bear with me here.
You may have seen me refer to a phase of animation known as "detailing" or "polish". There's a lot of small things involved in the detailing phase, but a big part of it is adding small variations to character motions, to give them more lifelike motion. Things like swaying gently as they stand, moving their heads as they talk, shifting their hands as they listen.
All of these micro-motions are accomplished through a process known as Jitter. The process is exactly what its name suggests, it's adding jitter, or small variations, to animation. Source Filmmaker has robust tools for generating and adding jitter. But it is a rather laborious task, because it has to be applied on a per-channel, per-bone, per-character, per-shot basis. That's a lot of per-'s.
The end result, though, is rather irrefutable.
On top of all that repetitive, precision selection, there is a lot of infrastructure that has to be put in place for jittering. Keyframes need to be manually counted out and placed along regular intervals, with adjustments made for existing keyframes that are near (but not on) the interval - have keyframes that are close together causes snappy animation that looks bad.
In total, it took me a bit over 2 days to jitter the Gaige video. And that's just an 8-minute video. Claire/Jill, recall, is over 22 minutes. That's looking at about a week of jittering. Just. Jittering.
Instead, I opted to go for the future-thinking approach. I've been jittering videos for years. It's pretty rote, pretty formulaic. And it's something I will be doing for years yet to come.
Do you know what we in the software engineering industry do to rote, formulaic tasks that will be repeated for long periods of time?
We automate them.
And so, that's what I have been doing this past week. I've already made significant strides with the tool, and have been putting it to practice on Claire/Jill, making tweaks and improvements to it as I come across the need to add them, all in the service of making my life easier and making video production faster.
With the tool in its current state, I have been able to take what would have taken me a solid two hours to jitter by hand, and gotten it jittered to virtually identical results in under fifteen minutes.
Of course, there is a trade-off in developing a tool alongside progressing a project. That tool takes time to produce, time that isn't put into the project. But just that short preview of things to come, I think, justifies the development time - I think it will pay for itself extremely quickly, in the time it will save me once the tool is done and I am able to mass-apply it to the Claire/Jill video, to say nothing of projects yet to come.
Right now, there's only one issues left to be resolved with the auto-jitter tool, and that is handling the case of jittering pre-existing animation. I've been using it on narrative sequences, which had no animation beyond lip-sync, because it's just talking and the specific character motions aren't important. And it works fantastically for those. I couldn't be prouder.
But now I'm getting into sequences where there are already animations, such as the sex scenes. And those need to be jittered too, but the tool doesn't currently account for it.
In theory it will be a simple solution.
There is a very large gap between theory and reality.
Jumping that gap is my next hurdle. And once it's cleared, progress on Claire/Jill is going to just skyrocket. I wouldn't be surprised if I manage to jitter the entire 22 minute videos in less than 2 full days of work. Hell, maybe even just a single day.
Wouldn't that be something? I'm giddy at the thought.
So yeah! That's where we're at. A new Witcher series with its first episode blazing through production, and a new tool to make Claire/Jill and future videos significantly less time-consuming and mind-numbingly dull to produce.
That's what we in the biz call "a pretty good week," if you don't mind my saying so!
Until next week, everyone!
Comments
Lyner
Thanks for the update! Looking forward to the Witcher project. Hope the tool works out.
Raptor_Demolished
Excellent update and absolutely huge appreciation for the Primordium news. Cannot wait to enjoy that. Keep up the great work!
