The cog, re-rendered: notes from week 19
We are on render four of the Hanseatic cog. For those keeping score at home: render one was our fault (the water was beautiful and completely wrong for the Elbe in autumn — too blue, too tropical, the North Sea does not do turquoise), render two was the hardware's fault (a fan died in the render rack at 3am and the node thermal-throttled through forty hours of frames), render three was nobody's fault (the museum changed a wall), and render four, as of this morning, is the one. Knock on the server rack.
Some things the case study won't say, because case studies are for clients and this is for whoever reads this: rope knots are a rabbit hole. There are three plausible ways to rig a shroud and our first pick came from a bad painting. Our historians caught it in week six, and I sulked about it for two days before conceding they were right, which they were, and the fixed rigging is better than the original would have been. If you ever work on a period-accurate ship: budget a week for knots. Nobody warns you about the knots.
The other thing: we test projection on the studio wall, which means a twelve-metre cog has been sailing through the office at 2× scale every evening this month. The neighbours have stopped asking. Yusuf has started conducting the score with a coffee spoon. It's a good project. See the finished thing at the museum from September — full case study here.