24

/

27

24

/

27

September, 17, 2026

Designing a HUD for Racing in the Skies

I've been designing a head-up display for a racing series that hasn't properly happened yet. Last week I finally showed it to someone who actually flies.

September, 17, 2026

Designing a HUD for Racing in the Skies

I've been designing a head-up display for a racing series that hasn't properly happened yet. Last week I finally showed it to someone who actually flies.

There's a bit in every project where you've got the work up on screen, the person on the other end of the call knows far more about the subject than you do, and you have to decide whether to talk them through it or just shut up and let them look - I decided to shut up.

The person looking was Zephatali Walsh, who flies for Airspeeder and also races FPV drones and flies crewed aircraft, which makes him one of a very small number of people anywhere who could tell me whether the thing I'd been drawing for months was any good.

What it is, and why I'm doing it

Airspeeder is a crewed electric racing series and the Mk4 is an octocopter, eight rotors in four coaxial pairs, and the bit that matters is that the pilot is sat inside it rather than stood on the ground operating remotely. The remote-piloted era proved the racing works and the crewed era turns it into a sport with actual people in it.

I've been round the Airspeeder factory and I was at the first public EXA race, none of which qualifies me to design a cockpit, but it does mean this sport has been rattling around my head for years. And what I kept coming back to was that nobody has built the bit I most wanted to see, which is the heads-up display.

So that became a project of its own inside the wider world-building piece I'm putting together. It's a display projected onto the visor of a helmet. I ended up with fifteen components sorted into three rough families: what the race is doing, what the aircraft is doing, and where you are in space.

The research, which turned out to be the easy bit

Before I drew anything I went off and read the human factors literature, most of it published by the US Army Aeromedical Research Laboratory, who have spent decades working out what happens when you put information in front of a pilot's eyes inside a machine that is doing its best to kill them.

Three things from all that reading changed what I drew. The first is binocular diplopia, which is the mess your brain makes when the two eyes are fed slightly misaligned images and it simply refuses to fuse them, and that's the reason so many fighter HUDs sit over one eye rather than both. The second is exit pupil tolerance, which is why a helmet that shifts by a millimetre stops working properly, and why an F-35 helmet gets moulded to exactly one head and apparently costs a small fortune. The third is field of view, and once you actually compare the numbers you realise this is where most science-fiction interfaces quietly fall apart (jarvis), because they fill the whole visor since it looks brilliant in a still frame, and your eyes just don't work like that.

Then I sketched it, and redrew it, and got to a version I was reasonably pleased with and quietly a bit nervous about, which is roughly the normal ratio.

/ A

The read test

So here's what Zeph saw, before I'd said a word about any of it…

The vehicle health cluster he read straight off the screen: ESC health, aircraft health, propeller health, R2 online, stabilised mode. The radar he got immediately, right down to calling a contact at seventy metres. Battery and boost he read correctly too, two laps of decent battery and nine minutes of flight left. All of which was great, and then it stopped going my way.

The altitude and vertical speed indicator he understood perfectly well and then told me they were on the wrong side of the display. My proximity and threat bar, which I'd spent a properly embarrassing amount of time on, he looked at and said "battery, maybe". And the track map, where I'd labelled the upcoming gate as G10, he read as a ten-G turn, because to a pilot G means G-force.

Two misreads out of six. Those two failures were worth the whole hour on their own, and what really got me is that both of them were things I'd already had a wobble about privately.

Where he was right and I was wrong

The big one was the motors. I had all eight rotors reporting RPM and temperature, which felt thorough and is in fact useless, because motors on a speeder sit in an enormous amount of airflow and are perfectly happy at full speed, and if one is running hot in all that cool air then something mechanical has already gone wrong and you've got bigger problems. The thing that actually matters is the ESC, the electronic speed controller, which takes raw battery power and turns it into the pulses that tell the motor what to do. As Zeph put it, the motor is a dumb coil and some magnets, and it's the ESC that decides how much torque and how fast, and RPM and current and temperature are all being read off it anyway.

So the number became ESC workload, which is amps being drawn as a percentage of maximum, and eighty to ninety per cent is high. That's a far more interesting thing to put in front of a pilot than a warning light as it just drinks the battery. So what you've got is a race management dial rather than a redline. It's the electric equivalent of fuel and tyres, which is the sort of thing this sport needs if people are going to shout at the telly about it.

A couple of smaller ones came out of the same conversation. Their display isn't calibrated against barometric pressure so it isn't altitude at all, it's height. And delta beats position, which he explained by telling me what he says to his own observer, which is that if somebody is gaining on him he wants to know about it, and if they're ten seconds back then be quiet and let him fly.

The bit I wasn't expecting

Somewhere in the middle of all this I started apologising for how busy it all was. Too much information, I said, I know, I'm going to strip it back.

And he stopped me, which I did not see coming. Cockpits look completely mad to the rest of us, but a trained pilot is formally assessed on the scan, which is the constant movement of the eye between instruments, and density out in the periphery is normal and expected and fine. Your focal point is about the size of a thumbnail held at arm's length and something like ninety per cent of your visual cortex is given over to it, so everything outside that is peripheral and can carry a lot more than a designer's instinct tells you it can.

Protect the centre, populate the edges. That's the rule I came away with, and I've spent most of my career decluttering interfaces on the assumption that less is always kinder to whoever has to use the thing.

The idea that beat mine

I'd gone into the call thinking the answer to overload was voice, so the pilot says remove altitude, remove position, and the display thins out around them.

His version is better and it's the thing I'll be building next. The AI co-pilot watches the pilot's psychological workload in real time and switches things off without being asked, because it knows this particular pilot and it knows what their stress looks like, since the two of them have trained together. So the pilot never manages the display at all, the display manages itself around the pilot, and that changes what the object even is. It stops being a layout and becomes something closer to a relationship. It also gives the co-pilot a proper part in the story rather than just a voice.

Underneath all of that sits the other thing we actually agreed on, which is that configurability is the feature. Let pilots pick what shows up and where it sits. He compared it to crosshair customisation in Counter-Strike, where the choosing is half the point and part of who you are.

What happens now

There are four states to design, so the standard ui rework, overtake mode, emergency message, and whatever the pilot ends up building for themselves.

Ótimo Studio is a brand and digital practice working with ventures and incumbents that intend to last. The Airspeeder work is independent creative direction, made with the cooperation of the team. If you're building something that doesn't exist yet and it needs to feel real before it does, come and have a chat.

Ian Howarth

/ B

Read More

Read More

September, 23, 2026

The legacy model of brand management is collapsing under the weight of modern volume. For lean marketing and design teams, the daily reality is a grueling, manual cycle of asset production where high-value talent is squandered on repetitive social posts and slide decks. To compensate for the chaos, designers are forced into the role of brand "police," fruitlessly defending guidelines that live in a static PDF nobody opens.

Solstice-article

14, September, 1026

Why the eight-hour brand is the most expensive thing you will ever build

Tradelocker-article

September, 23, 2026

The legacy model of brand management is collapsing under the weight of modern volume. For lean marketing and design teams, the daily reality is a grueling, manual cycle of asset production where high-value talent is squandered on repetitive social posts and slide decks. To compensate for the chaos, designers are forced into the role of brand "police," fruitlessly defending guidelines that live in a static PDF nobody opens.

Ótimo

© 2018-26

© 20

26

ÓTimo IS a PURPLE PATCH LIMITED COMPANY. All rights reserved.

Ótimo

© 2018-26

© 20

26

ÓTimo IS a PURPLE PATCH LIMITED COMPANY. All rights reserved.

Ótimo

© 2018-26

© 20

26

ÓTimo IS a PURPLE PATCH LIMITED COMPANY. All rights reserved.

MALTA, MT
MALTA, MT