I have been building a radio telescope facility in Unreal Engine. Not for any other reason than because I want to go there.
The facility is in New Mexico. Three steerable parabolic dishes, a control room full of consoles, a diesel generator that needs refueling, a bunk with a coffee machine and a shortwave radio. Outside, snow and sky. Inside, the hum of cooling systems and the phosphor glow of waterfall displays scrolling through frequencies.
The project is called Strange Sounds, and it started as a game design document. Somewhere along the way it became something else. A place I am building because the real version does not exist in any form I can access, and I want to sit in a control room at 2am and listen to what the sky is doing.
Why a radio telescope
Most people think of astronomy as visual. Hubble images. James Webb infrared. The pretty pictures. Radio astronomy is the other side. It is listening, not looking. The universe is noisy at radio frequencies. Pulsars tick like clocks. Jupiter crackles with electromagnetic storms. The Sun hisses. Fast Radio Bursts arrive from billions of light-years away, last milliseconds, and nobody knows what causes them.
The instruments that detect these signals are not telescopes in any recognizable sense. They are arrays of parabolic dishes connected to receivers, amplifiers, and signal processors. The data comes out as spectrograms: time on one axis, frequency on the other, intensity as color. Reading a spectrogram is like reading weather. You learn the patterns of noise, and then you learn to see the things that do not belong.
I find this deeply calming. Not the science specifically, but the practice. The act of pointing a dish at a patch of sky, tuning a receiver, watching a waterfall display, and waiting. Most of the time, nothing. Occasionally, something. The ratio is the point.
What I am actually building
The facility has eleven console stations, each modeled on real equipment. The Observatory Planning Console uses RA/Dec coordinates and Gantt timelines to schedule sky sweeps. The Antenna Control Console has azimuth and elevation wheels with mechanical inertia and wind drift. The Receiver Tuning Console lets you choose bandwidth, gain, and local oscillator frequency, with a tradeoff: wider bandwidth captures more signal types but raises the noise floor.
The signal processing chain is real. Gaussian noise background plus galactic interference plus weather effects. Natural signals modeled on actual pulsar profiles and solar burst characteristics. The demodulation pipeline does AM, FM, SSB, and digital modes. You can watch a constellation diagram converge as you dial in the symbol timing.
I built all of this because I wanted the controls to feel like controls. Not game abstractions. Actual instruments with actual tradeoffs. The kind of decisions a real operator at a real facility makes at 3am when something unexpected appears on the waterfall.
The virtual vacation part
I used to travel for work constantly. Airports, hotels, but now with four kids the places I actually want to be are remote, quiet, and full of instruments. The Atacama Desert. The NRAO in New Mexico. Jodrell Bank in Cheshire. The Australian Outback where the Square Kilometre Array is going up.
I cannot live at any of these places. But I can build one. Not a replica. A composite. The control room layout borrowed from photographs of real facilities. The console designs referenced from actual antenna control systems and SDR software. The exterior from satellite imagery of high-desert research stations.
The idea is not simulation for its own sake. It is that there is a specific quality of attention that comes from operating instruments in a quiet place, and I wanted to see if that quality transfers to a virtual environment. Whether sitting in a virtual control room, tracking a signal across the sky, adjusting a receiver to pull a weak pulse out of the noise floor, produces the same focused calm as the real thing.
Early testing suggests it does. Not identically. But close enough. Building systems, learning things, scanning the sky for new frequencies without having to get approval from the housing association to mount a dish on my roof.
Space sounds
The part that surprised me is how interesting the audio layer became. Radio astronomy data is not audio. It is electromagnetic radiation at frequencies far below or above human hearing. But you can shift it into audible range, and when you do, the universe sounds like something.
Pulsars sound like ticking. Some fast, some slow, some with a swing rhythm like a jazz brush on a snare. Jupiter’s magnetosphere sounds like ocean waves crashing through a broken speaker. Solar bursts sound like someone dragging furniture across a wooden floor in the apartment upstairs. The cosmic microwave background, when sonified, sounds like rain on a tent.
I have been collecting these. Not recordings of actual radio telescope data, though those exist and are publicly available. Procedural recreations that behave like the real signals: frequency drift, amplitude variation, the way a pulsar’s period changes as the dish tracks it across the sky and the Doppler shift accumulates.
The game generates these in real time. The waterfall display and the audio are the same data, rendered two ways. You see the signal and hear it simultaneously. When you lock the tracking and the SNR climbs above threshold, there is a moment where the sound clarifies out of static, and it is genuinely moving. Every time.
What this is for
I do not know if this becomes a commercial product. The game design document exists. The vertical slice milestone is planned.
But that is not why I am building it. I am building it because the act of constructing a place I want to visit has turned out to be almost as good as visiting it. The research required to make the instruments authentic has taught me more about radio astronomy than any course would have. The nightly sessions tuning virtual receivers and watching virtual spectrograms have become the quietest part of my week. A small virtual holiday.