OK, let’s get a few things out of the way:
- I’m a nerd. I mean a proper, hard-core nerd, who cares about reactor volumes and power to weight, internal consistency to long forgotten lore, armor grazing angles. Stuff like that, stuff that movie audiences don’t really care about. I get that I’m different, and that’s it’s absolutely ok not to pander to me or to people like me in making movies.
- I’ve been doing CGI SW ships for a long time now. I did a CGI Executor in 2001, and an ISD in 2003, the old version of which basically everyone uses (including for commerical projects…mumble mumble). I’ve done dozens of original designs, some of them for official sourcebooks (Essential Guide to Warfare). I think I have a decent grasp of what a star destroyer-type ship ought to look like, and where to go with it. Go to the rest of my website (https://fractalsponge.net) and see for yourself.
- The Force Awakens was fine. Sure it was a reshoot of ANH with nice CGI. Sure Rey’s character was a Mary Sue. Sure the villains didn’t stand up to Palpatine and Vader. Sure Starkiller Base was wanky. But it was visually beautiful and scratched a nostalgia itch and had a charming cast of heroes and just plain dared to make something cinematically new in Star Wars, however incrementally. The ship designs weren’t instant classics like in ANH but they were fine, some were even awesome (Kylo’s shuttle for instance). I have zero problems with not using original trilogy designs when making a new trilogy.
- I know it’s too late to change anything. By the time we’re seeing renders like this, the production process is going to be pretty locked in.
So, we’re here a few years after TFA and some designs from The Last Jedi are coming out. I’m going to focus on the New Order Dreadnought. It’s 8km long.
Here’s my art direction reaction to the design: it sucks. But rather than leave it at that, I’m going to go into a lot of detail as to why I think it sucks. I mean a LOT of detail. Turn back now if you think you can’t handle it.
——
You’ve been told to make a giant spaceship, a huge one, a city in space with engines. But no one’s ever built one in real life, so your sadly Earth-bound audience has no idea what it really should look like. There’s opportunity there because there aren’t a lot of pre-conceptions about what it ought to look like, right? But let’s be honest, this is Star Wars we are talking about. People absolutely DO have expectations regardless of how the actual future turns out. It’s McQuarrie and Cantwell and that classic “oh shit” moment when we all saw Devastator fly overhead chasing the Tantive IV for the first time. No, it doesn’t mean Imperial Star Destroyers and X-Wings forever. It categorically does not mean that, please Disney do not do it. But those old ships and many new designs in the prequels and EU had a shared design ethos about them, used and functional looking despite just being tons of Tamiya plastic model kits glued onto styrene and fibreglass. Complex, full of character and screaming that there was a history behind every little box and turret. Now we’ve all heard about how strong designs should be captured in a few strokes. Yeah, design integrity is good like that. But IT DOESN’T STOP THERE. You can’t draw 3 splines, extrude them, and call it a day when building ships for Star Wars.
Here’s a practical example. Assuming you know nothing about ships, what ship looks bigger?
Now if you knew something about ships, you’d know HIJMS Fuso (bottom) is bigger than USS Zumwalt (top). But you’d also know that Zumwalt is >90% of Fuso’s length. Does that surprise you? It should, because simple shapes look a lot smaller than bigger shapes. Turns out surface complexity, and obvious surface complexity, makes a ship look bigger. Take this ship, my design:
It looks very similar to the FO Dreadnought. Slab bridge module and a big simple triangular plan. But the fact that things jut outward from the model makes it look complicated. The ships might look about the same size. But this thing is 600m, and scaled for it, vs 8km for the FO ship. That ship is the size of a damn city, but does it look like one? No.
Big constructs, in most settings, are built. They are not carved from monolithic blocks of material (except for structures in universes where that is explicitly true). A bigger ship generally has structures on it from smaller ships and that helps to build the sense of scale. Like how Resurgent added more layers of terraces to the ISD, so that even without the bridge module as a direct comparison, it looked pretty big and looked plausibly like a design successor from the ISD. That wasn’t automatically bestowed upon it by being a triangle.
Here’s another example from my stuff:
It’s <4km. Still smaller than that ILM ship. Does it look bigger? I think so! The surface is broken up by obvious structures, and the profile is irregular. It looks like it may have been *built*, rather than just injection moulded and painted splotchy grey.
Now there’s a ton of random stuff hanging onto the top and brim of the ILM ship. Great. But it doesn’t matter. Because when you see the whole ship in an establishing shot, no one can count windows or see how many little turrets there are – they are just bumps. And when they are just bumps without any obvious granularity to them, their scale becomes impossible to guess at, and it melts into background noise. If it does that, why bother to model the little nurnies and surface bits at all then? It’s a waste of polygons and computational power.
I’ve made the exact same mistake before:
This ship is supposed to be big, and has some nice detail on it. But the problem is the profile wasn’t broken up enough. It’s plausibly any of a huge variety of sizes, because there aren’t enough scaling cues, and the complexity that is there doesn’t break up a fairly monotonous outline enough. But I am some schmuck on the internet. The people making The Last Jedi are the designated successors of Mcquarrie. They should be doing better.