Wednesday, June 13th, 2018:
At the beginning of the day, I started off with Brandon again until more people trickled in. We began to do polish work on the shot. He wanted to wait to do the hands, and focus more on the flipping action to get that down on the fight line, and timing to be right for each pose in it.
We started with a quick pass of the hips and torso just to see that the curves aren't funky in the translates and rotates - this is to make sure all hits well on the fight line before going into other body parts. This is important, because when doing another pass on the curves, sometimes the stuff that looks like junk is actually data that you need.
After the extra look over on the hips and torso, we then baked down the animation, and started moving on to the nitty-gritty.
Brandon and I first looked at the translations, because he says you can see a wall hit in the translations way quicker than in the rotations, so focus on those to see if there is any hard pops. Again, in doing all of this, we are comparing the work being done to that of the original data, because we want to be as true to that data as possible.
After polishing for a while, the executive producer, Selma Edelman, asked me to start shadowing the post-production coordinator, Kelly Huffine, to see the initial setup for big projects like this.
While sitting with Kelly, I learned their current client will send them raw data that their company has captured, and she receives it in the form of FBXs and spreadsheets. The spreadsheets include the names of all of these files, and then the descriptions of the desired outcome for the shots (the actual actions the character needs to make). She then brings those into Shotgun (the folder, file operating system that they use here at H.O.M. to create master projects), and sets up the contract (naming conventions) and sequences (character setups). These also need to showcase task templates, which seems like more of the pipeline of what steps still need to be taken on this shot (camera, layout, in progress anim, etc.)
This is then sent out to the mocap tech animator to begin retargeting and then to the lead animator of the project to assign shots and have things in correct folders (characters, props, stances, etc.)
After learning about Kelly's doings, we went into a Skype session meeting with this video game client of theirs to discuss progress, questions, and decisions on certain parts of the cinematic. In the meeting, it was a lot of discussion and terminology that I didn't understand simply because I don't know the 'code' names of everything yet, but most of it, I was surprised to understand. They are having trouble using this company's rig for the mocap data. They have the facial data on it, but have to do a Maya transfer to have the rest of the body data applied to this specific rig correctly. There were also some questions about the rig system and constraints/rigging when that transfers from Maya's system to MotionBuilder, so I was glad to know that even the industry has rigging problems and transfer issues when blending between two softwares.
From here, I spent the rest of the day with Troy Reynolds, who runs the mocap stages here. We spent about 4 hours together going through their system, calibration, and running a mocap session, that way for future shoots, I am able to help with the system performance and capture. Tomorrow we are demo-ing to a school group, and then on Friday we are doing a full on capture session for a third game we are working on.
CALIBRATION: We started with doing the normal set up. They have 88 cameras on their main stage, so masking takes a bit longer during the calibration process. Troy said that sometimes, he will actually take the cameras, and mask them in chunks, otherwise the process could take up to twenty minutes. From there we did the wand wave - of which they have a bigger wand since their space is wider, and needs more volume coverage. Then we did the origin set up, which is different. Instead of placing the wand down at the center origin, they actually have a pyramidal object that Vicon supposedly used to make before the wand to set for the origin, and it works better in their space. And then finally, we set the floor plane (same way with about 5-10 markers on the ground plane). Below are some pictures of the pyramidal device for the origin, as well as comparison size of the wand here to the normal wand we use at Ohio State.
BLADE V. SHOGUN: Here at H.O.M., they have both Blade and Shogun. But just like what we have found at OSU, Shogun is still so new that there are a lot of bugs that need fixing. So until those issues are resolved, Blade is the primary software being used.
VICON: Troy told me that House of Moves and Vicon actually were run out of the same building here in L.A. years ago, so he really used to work for them beforehand. Now, he has a meeting with them once a week, which he said I can now come to, to discuss whatever glitches he finds in the software - especially since we have one of the largest combination of cameras on a single stage, which calls for more attention.
FACEWARE: There are two Faceware helmets that they have here for capturing. This is something I look forward to messing with, simply because I have done body capture, and I have done face capture, but never had the two going at once, so I am eager to learn the process of how it all comes together in a real production pipeline. Troy told me that they also have Cara here that hasn't been used much, but definitely something that we can play with. They also have Dynamixyz, but that doesn't have any hardware that they have, so they have just been using Faceware's helmet that they own with it.
GIGANETS: Here, because the system is so big, they have about 9-10 giganets running for all 88 cameras.
SUITS/MARKERS: Because there is clientele and talent that comes in for projects, there is a 'client' area with dressing rooms, relaxation areas, and just an overall space for comfort while shoots are in progress. The suits are basically the same as ours, but color coordinated so that reference video can be used to distinguish which actors wore what colors and what actions they perform.
C.O. AND Q.A.: Camera Operators and Quality Assurance positions are needed whenever there is a shoot occurring .
DAILY HOURS: 9:00 a.m. - 6:00 p.m. (9 hours)
CONSECUTIVE HOURS: (27 hours)