The sky was briefly beautiful this afternoon. I looked up from the screen to our east-facing balcony, holding my neck at an unfamiliar angle and couldn’t really think about anything except how rich the clouds were and how the gradient of blue to dirty gold at the horizon looked like the backdrop of a massive set I’d accidentally wondered into.

Normally this blog and the digital sketch are mostly done by the weekend, I chuck in things here and there over the beginning of the week, give it a quick glance over in the morning and then post. This time I’m cutting it pretty close to the wire, staying up late the night before. This may be because I kept fiddling with the sketch; I’m not happy with it. It feels too cliched, there’s no focus, the lighting is a bit janky. Interiors are hard. So I chucked some volumetrics in and so infinitely multiplied the rendering time which means it was only finished about 15 minutes ago. I also didn’t get to really play with the new global illumination addon because I realised I didn’t know enough about Blender‘s existing global illumination setup, hence the new learning curve, hence the fiddling, hence the lateness.

We can Simulate it For You Wholesale

I’ve been chewing over a thought that popped up in some writing a few weeks back: It feels like realism is entering a really contested space. There are so many different versions: cinematic, animatic, climatic, natural, mathematical. All of these have their own versions of ‘realism.’ Even then, within each of these larger version of realism are contests based on the limits of the technical frame through which most folks experience these realisms. I want to try and unpick this with examples of two at the boundaries of video game-esque simulation: Microsoft’s new Flight Simulator (which I haven’t and won’t ever use) and the interactive cycling ‘simulator’ Zwift (which I do use).

Both of these software, though based on video game engines and technology are edge-case video games. They aim to simulate certain effects of a real-world experience, of flight and cycling to a degree of realism. It’s important to note that the briefs are different: Flight Simulator does fall into the fantasies of video games in the sense that it aims to connect (most) of its ‘players’ with an experience they can never have – flying a plane. While Zwift doesn’t and can’t seek to simulate cycling, only to give a passably entertaining experience of cycling on the spot in your flat. And, unlike Flight Simulator everyone who uses Zwift is necessarily also a cyclist – because you need a bike.

What do I mean by a contested space? Should you ask rhetorical questions when writing? Well, most desktop computers are now very good at rendering good quality graphics relatively easy, both thanks to advances in the software and the hardware and standardisation of certain technical processes that make them more interoperable. However, there has to be some concessions, there’s only so much that can be simulated and so the world has to be built in response to imagined use-cases. So what decisions are made when someone, creating a world, defines the boundaries of realism? How do you predict the actions of an actor or individual in your simulation such that all the affordances of the world appear as realistic, and not as simulation? Let’s briefly explore a corollary from Blender:

Given a limited technical frame, limited either by money, memory, power or speed, decisions have to be made about what is primarily constitutive of reality in constructing the simulation and what is secondary.

For instance, Blender‘s ocean modifier tool is remarkably good at really quickly simulating what appears to be the surface of a body of water. This might otherwise take hours of messing around with procedural displacement, but in a few clicks a relatively new can make an ocean, pond or lake surface and animate it, even simulate foam and spray. (I know this because I do it in day one of my Blender 101 class.) The good folks at Blender (which, remember, is produced for free by a foundation) decided that this was a tool users would need to improve their workflow, but crucially they also decided that this tool would mostly be used at a scale 10 – 100 meters. If you attempt to render a scene any closer than 10 meters-ish to the result then the geometry and lack of surface detail becomes obvious. You can fix it up with some extra stuff but that’s not part of the drag-and-drop modifier that Blender have designed. Any further out than about 100 meters and you’re forced to tile it, resulting in an obvious repetitive pattern. Again, this is fixable but requires an extra bit of know-how and experience.

Blender had to make a judgement call, given a limited technical frame of the average user’s computing power, technical skill and preferential aesthetic to deploy use these particular algorithms to make this particular modifier that works in these particular cases. Given a limiteless amount of computing power, this would not be a problem, but realism has been zoned to 10-100 meters. Blender have circumscribed a technology based on certain expectations about its use. To draw on my favourite paraphrased definition of technology from Alfred Gell, it is a tool (the computer, Blender), the knowledge on how to use the tool (which has been made as simple as possible by a drag-and-drop modifier) and the social necessity of its use which is limited to 10-100 meters.

(Don’t even ask about horizons. There’s a reason you don’t do horizons.)

So, back to Flight Simulator and Zwift. Nearly universaly hysterically positively reviews of the new Flight Simulator have pointed to its astounding accuracy both of the experience and attention to detail. This has been achieved, as previously noted, by Microsoft plugging together various bits of its ecosystem like mapping, weather, 3D scans and advances in procedural rendering into Flight Simulator to produce what appears to be a vast, realistic world.

I think, with this type of thing we sit at the edge of an interesting sea-change where it’s in fact easier to conceive of the game as a god-mode tour of Microsoft’s GIS data trove as opposed to a simple flight sim. Similar to the way video-game publisher and faceless world-eaters Electronic Arts have become the de facto gatekeepers of all football data, real or virtual such that it’s easier to see them as data brokers than game developers. But, even with all this global data the trick of realism only works at the scale within which you were intended to interact with it.

For instance, it uses Bing maps (lol) to generate the whole world and then extrudes building and landscapes from available data. However, limited by its technical frame (Internet connection speeds, desktop computation) it makes best-guesses about how to generate these 3D forms and so you end up with Buckingham Palace rendered as a drab office building. In the form of interaction intended by the designers, these oversights are forgivable being that you might for instance decide to cross the whole of Eurasia in one session and as a gestalt experience, the whole thing needs to be accurate enough. The social necessity (believable, global flight) makes use of a an imperfect tool.

Briefly then, Zwift also offers a rendering of Buckingham Palace, hence the useful parallel. Again, it’s worth pointing out that the designers have no intention of suspending disbelief; the Zwift maps are playgrounds – everything is out-sized, brightly coloured and simplistic. Zwift took a punt on the chance that most users would either not be Londoners or be not familiar enough that they would believe it by landmarks. London in Zwift is a trope of London, and Buckingham Palace is a trope of Buckingham Palace; all stone and imposing but this isn’t streamed, it’s all on your hard drive so the models are low-resolution with simple geometry and recycled assets. In fact, the quality of the palace model in Zwift is comparable with the one in Flight Simulator for geometric detail. The realism of Zwift has nothing to do with the visuals, it’s all in the relentless stream of data and feeling in the legs. It’s never going to convince you that you aren’t in your living room and it doens’t want to. It just needs to hit the tropes and symbols of London to make it cartoon fun.

Flight Simulator‘s glitches are a counter-factual rendering of reality in which the expectations are well-defined (Buckingham Palace is a big stone, classical building) but the reality (it is an office block from like, Swindon) conflicts. In Zwift, the expectations are cartoonishly reduced and out-sized, enough to trigger acknowledgement but not enough to suspend disbelief.

Really I’ve been writing a lot about expectations and reality so this stuff is jsut rattling around up there.

Recents (Today)

Ok! If you’re reading this then there’s still time! At 1030 BST (today, which is 15 September) I’m hosting David King, Annie Goh and Deborah Tchoudjinoff in a panel on their practices in world-building and at 1600 BSY I’m going to disassemble some world-building with Christopher Lutterodt-Quarcoo and Georgina Voss. What are you doing? Get over to Eventbrite and sign on up, links are in this paragraph. It’ll be like a big party.

Short Stuff

• I almost made the whole post about this ↓ but I got distracted by renders of Buckingham Palace. Anyway, when Epic Games aren’t busy trying to disassemble Apple and Google’s app gatekeeping monopoly they’re also making significant leaps in special effects and developing the Unreal engine (of ‘Holy shit! Have you seen the Unreal 5 demo?‘ x 16/week fame), the world’s most successful video game engine. This totally slipped me by but it was used on The Mandalorian: a massive LED screen with a Unreal real-time rendering on it, on a sound stage. The death of green-screen? They’ve basically collapsed a massive part of the post-production chain. This is super cool but very upsetting because of all the Manovich and Levitt theory I’ve read that draws on the aesthetics of post-production, which may not be around much longer. Oh well.

• You ever get the sense that Facebook still hasn’t figured out what it’s for? That the pomp and bombast is wearing thin? Like Robespierre toward the end they’ve both failed to bring the promised land and keep the advertisers happy. Sort of easy to forget that it’s a dying empire. The latest 20 Minutes Into The Future bulletin concerns the world after Facebook. The signs of ageing are already there; it’s slow to respond to changes and threats, can only survive by copying or stealing and is hitting an upper limit of the viability of an out-dated model. (Literally. The number of dead users is set to surpass live users in the 2060s.) But what happens to cultural memory and heritage when one private company holds all of it?
• In my apparently weekly reading on Disco Elysium (why haven’t you played it yet? Its only like 20 hours, take the weekend my friend.) I was frustrated by this article on Vice about how a lack of sincerity holds it back from ‘being the game about Communism it could have been.’ I’m not sure I agree with all of it. The author seems angry that the games’ steer towards characters that self-deceive and obfuscate their feelings makes the fabled revolution impossible. I agree that a better world requires sincerity and honesty but calling on the main character to get over his own very real internal emotional turmoil because his ‘solidarity should be elsewhere’ is a flawed political strategy that we should not be inspired by. If you read any interviews with the creators (who are avowed communists) they highlight self-deprecation as a starting point in levelling the playing field of political discourse. And boy, does Disco Elysium self-deprecate.
• Here’s Venkatash Rao on the pitfalls of charismatic technology. Golly he’s smart.
• Here’s beloved collaborator, colleague and friend Charley Peters on the Art Fictions podcast. It’s all structured around the Yellow Wallpaper, which I’ve never read but there’s so many great things there; ‘might as well talk of a female liver.’ I listened to it while writing this (me n Wes get a shout out).
• Since this whole thing has been about simulation. Go to this, which looks at the other side of the whole thing.
• I had more short stuff including something related to the new Epic/Unreal tech but once again I am using A Content Strategy and saving it for next week.

Those short stuffs weren’t very short, I appreciate that. But, you know, once I start I get too lazy to stop. Anyway, I love you as always, I’ll write you next week.