Thread by @Duecenage

LOOL! Hope you enjoy my novella! It depends on complexity of the scene. A lot of times, the boards are clear and we're able to just truck along and elevate it. Other times the boards and assets don't quite line up, that's were I'll start to "translate" things for my crew.
(1/25)

https://twitter.com/jj_conway/status/1690453844894035968

In some cases, that means, "Let's stage everything over by the mailbox. Someone interacts with it for a shot, so it's technically the most important element to the 3D staging" And then we're basically able to follow the boards for the characters, and the BG will do what it does.

And then for an action sequences, there is a whole bunch of moving parts. Plus, when the boards and assets don't match up I get more granular with my translations. I'll be talking specifically about the Episode 14 chase sequence from here on forward.

I start by assessing my crew and picking someone experienced enough that I only need to direct them on what to do, and not how to do it. It's a bit blunt, but TV animation goes fast, I find it keeps things running smoothly. (This is only for the toughest of the tough sequences)

Then I get them to watch the animatic, start putting assets together and exploring cameras and space (Take the longest shot, and see what speed Bee travels at for it) At the same time I watch the sequence on loop a dozen times. Taking notes, playing out the action in my head.

I'll also load up the assets in Maya and start exploring, looking for spots that match what is in the boards. We didn't have a drag strip in the racetrack, but we had straightaways, combined with a lamppost to go around, it kinda fit to match the dragrace section of the sequence.

I also start mapping. Even though its chaotic, I still want things to flow logically when looking top down at everything. Then I can relay that to my animator "from this shot to this shot, they're driving this way" (Here's the initial map, I drew a cleaner one for my animator)

This is also to help ensure the lighting remains intact for rendering. I don't want characters to suddenly be turned 180 degrees from shot to shot, which would flip their shadows and rims. A few degrees off is usually fine(some Art Directors might beg to differ, lol)

Then after that we previz. We usually don't have time to previz, we just go straight to layout. But I wanted to know early if I was going to need to ask for a major set change for this race track. So my animator starts doing some shots, I start doing others.

They send tests, I load their shots and see if there is any way for me to push things beyond the surface level limits. I get a little more hands on here cause I've done lots of talks with CG Supes and department heads, so I think I know how close to the line I can push things.

"Great power, great responsibility. All that jazz!" I only show these tricks to less experienced animators if I know it's a faster solution. I'm a "5 minute fix" instead of a "4 hour redo" director. I just don't want them to use it all willynilly, it might hurt other departments.

And in that same vein, that's why I previz'd. If I could use what we have and get even 75% as dynamic as the boards, then it's worth doing. A new asset would've made things easier for us, but maybe harder for them, and I think it was a net time save overall if we(anim) solved it.

Since my lead was familiar with the sequence during previz, I assigned him the layout. And then noticed that it would put him over his budget. So I took on the excess shots.

I tend to grab the "easy but tricky shots". My animators are better animators than me these days, but my layout is clean, concise and I'm damn good at cameras. Some of my alltime favourite sequences are ones where I did the layout, and another animator animated them.

For this sequence in particular, there was a shot that lived the clearest in my head. I had submitted a request to change it and was allowed to show in layout what I wanted to do. Would've taken forever to describe how I was seeing it, so I chose to show it myself.

So, since my lead and I did the layout, if I got extra help, these shots were in a good enough place to have anyone work on them. And then once animation was done, I went in a set decorated everything.

In layout only the pieces that were interacted with were there. So if Bee drove behind a flipped SUV, or jumped over the crashed semitruck. those were in layout. And then once everything was flowing nicely, I did a continuity pass. I put together a list to help track things.

It was near the end of layout where I had the realization to do it this way. That it would be a lot easier to set decorate after the fact. That way we could nudge objects to help hide the cheating we had done to extend the size of the racetrack.

I was constraining SUVs to lampposts and sliding those to wherever made sense and looked good in the BG. I forgot to maintain where the lights of lampposts were pointing(I would just translate and rotate to what made the SUVs look best), but I got lucky there.

I did the same thing in episode 21 with regards to spotlights coming from off screen. If Thrash was 30 feet away and looking in Jawbreaker's direction, then Jawbreaker should be getting some light from Thrash chest lamp, even if he wasn't in the shot.

The part that made it all click for me in ep14, was when I started playing with the focal lengths. I picked shot, set the focal length to 11mm and then suddenly Bee looked exactly like he did in those storyboards. A light shone down on me, and I relayed all that to my lead.

We made the changes when going from previz to layout. I'd already gone wider than usual. 28mm was the norm, with the odd 2 to help push depth. My personal focal length values are(in mm): 20, 24, 35, 50, 70 for acting, and 18, 20, 24, 35 for action. Sometimes a 15 for some spice.

For the racetrack, 18mm was where we started, and then we went to 15 for the more dynamic shots. But then we changed to 15 for most shots, and the odd 11 to really, really push things. It blew me away how much the opened up the race track.

We couldn't use those FLs when looking across from one straight to the other, cause we could basically see end to end of the race track, and it left us no room to fake the track size. So we used them when lookinhg down the long way.

I think I've rambled long enough(and Twitter seems to agree!) The funny thing this, this just how my brain works. Typing this out required a more concentrated effort than actually doing it. I like to just get my feet wet and figure things out along the way.

@threadreaderapp unroll