Applying SpaceX’s production process to the development of memorable experiences

I’ve been so busy lately that I haven’t had time to do a write up of our first prototype of Karl-Heinz StockHausen’s Kugealauditorium, a project I described in a previous post. Anyhow- better late than never, so here’s an update on KA2.

Recap of the project

For Kugelauditorium, our plan is to develop an open-source recipe for spherical concert halls that are cheap and easily deployable. We’d like to create a universal instrument that can provide for its users expansive psychonautical journies capable of vehicling them to the outer reaches of human consciousness, or, if not, then at least give them an engaging audiovisual experience. The provisional dream is “a Kugelauditorium for every garden”.

For that to be the case, we need a very simple and cheap design that anyone can knock together in no more than an hour or two, that’s sufficiently magnificent for human consciousness that it is more highly valued than, say, a lawn. In other words, something an order of magnitude or two simpler than any present-day equivalent immersive sound experience, but at least as compelling.

To develop such an ambitious recipe, we’re engaging in a rapid prototyping methodology loosely inspired by the Starship development system currently being executed by SpaceX at Boca Chica, Texas.

Obviously our project has much more upside for the future of human consciousness, but SpaceX’s work bakes in a few pretty robust insights about project management, decision-making and building towards ambitious future scenarios in a speedy way.

Their method in a nutshell is to optimise for the number of iterations in the development process. So rather than even trying to get everything right first time, they build speedily and iteratively and holistically. The genius of this method is various:

  1. First, it forces simplicity. With the luxury of time, one can over complicate anything, and one will. By insisting on an improved prototype each month, we force selection of the simplest thing that will work. Relatedly, because one has to spread the budget over many prototypes, you have to select for cheapness.
  2. Second, by aiming for a holistic working prototype each month, one has to think everything through together- you can’t go too deep on electronics or speakers for example, and completely forget about the structure. So one always remains tied to the whole end user experience, and everything has to be thought through together- and this means that solutions at the intersection of different parts of the system emerge.
  3. Third, since everything has to be designed so that someone else can build the next version somewhere next month, the whole is optimised for easy reproducibility and transparency of design. Since the goal is a widely deployable open-source Kugelauditorium design, this is fit for purpose.


We began the project in April, and aimed for the first prototype in May. This interestingly led to an initial budget estimate of $50k for a KA being compressed down to a budget of about $12000- the merits of a tight time scale eviscerating all indulgence.

Here’s how we did it:


We built the dome using a very simple compact kit provided by Build with Hubs. They provided just the joins between struts, and the struts themselves we got from a hardware store for $150. They came in 50 odd 240cm round poles, and we cut them at a single point in the correct ratio to waste no wood (1179 vs 1221mm for short and long segments, if you’re interested). There’s a nifty thing on their site that helped calculate the ratio. Using two spherical domes, placed one on top of each other, we could get a fully spherical dome for $400. 

Timelapse of dome construction
The two hemi-domes joined together


Next, for speakers, we got two cheapo 5.1 surround sound speaker systems that took care of amplification and presented a total of 10 channels to address. That allowed a 360-degree sound experience, though only 2D- as the speakers were all arranged around the middle rung. On the software side, we used standard tools . All of this part was done by Tomek Smilok.


To simplify things on the light side of things, which was not our main focus, we used sound-reactive LED strips, so as sound travelled round lights were activated with the moving sound (this saved requiring programming the light by itself) and provided a glimpse of how the light can in future be super-additive to the sound. This was the work of Iannis Bardakos and friends, and involved lost of soldering.


Tomek created an improvised symphony entitled ‘backpain’. Of course, till you have a Kugelauditorium in your garden, you won’t have the faintest idea of the magnificence of the experience but here’s a sense of it.


Phenomenologically it felt relatively spectacular, definitely tremendously immersive and engaging.

The faces tell the story of the wondrous phenomenological adventure
KA2 at a distance.

That was Kugelauditorium 2, (KA2) and it was pretty epic. The experience of the surround sound was great -one felt held as if a spirit in space, and it was unexpectedly comforting- I suspect the final instrument may have some pretty potent psychotherapeutic uses. Visually, the full sphere looks more than twice as cool as the generic hemi-dome, justifying the indulgence by itself, and we were able to get a glimpse of the full experience. Undoubtedly, this represents an instrument that every garden in the world should possess.

Total cost was $1200, and it took a Saturday to knock together.

Areas to focus on in KA3

We figure that a dozen or so iterations will be enough to get to something truly masterful, so we have no time lose in keeping iterating. The next prototype will be built in London in July, with the following one likely in August by Californian sound artist Sarah Stevenson.

The principal limitations in KA2 were:

  • The sound was limited to ten channels and was only 2-d (360-degree surround but on just one plane)
  • The positioning of people in the structure was at the bottom not in the middle
  • We didn’t really have time to experiment with the music.

The key issue we’re addressing in KA3 is the first, making the sounds fully three dimensional and over many more channels, including the capacity to map the sounds to that space in software easily. We’ll also aim to develop the structure so that a person can be suspended in the middle.


KA3 spec: our main project here is to make the sound fully 3-D. So we’ll use the same basic dome structure, but use stronger wood to create the possibility of a single person being suspended right in the middle to enjoy the fully three-dimensional sound immersion.

We don’t know how many individual sound channels are required for fully-3D maximal precision of audition, but because the dome structure has 42 joins (from top to bottom at the different layers: 1/5/10/10/10/5/1) and because is 42 is more than 10 and because audition isn’t terribly spatially accurate, we’re going on the basis that 42 will more than cover it, and 41 in fact since there is no lowest point. That we guess will be perceptually equivalent to having 1000 speakers in terms of spatial precision (we may be wrong about this, of course).

For simplicity for the next version, we’re cutting this 42 down since 32-channel audio is relatively easier to manage from a hardware perspective, so we’ll aim for that, provisionally with 1:5:5:10:5:5:1 at the different levels of the dome (or perhaps 1:4:6:10:6:4:1), and a subwoofer at the bottom to make 32.

The basic diagram of the architecture is like this:

So our complete recipe for the kit (not including lights) for the KA3 will be this:

  • Two geodesic dome kits from Built with Hubs (it’s cheaper to get your own sticks, but these kits including sticks are very convenient- (2* £275)
  • two multi channel DACs, specifically the Cymatics Audio live player LP16 (£215*2)- this was the most exciting piece of kit to discover, as 32-channel DACs normally set you back $3000.

So in total we get to £2300 before cost of any lights we may add. Not bad, but still touching $3000 and twice as expensive as the first prototype. We will need to build efficiencies into the prototyping process, but we should still get 12 prototypes in for the original Bitcoin which was swapped for the dream of a Kugelauditorium.

With this recipe, we can get to a fully spatialised 3-D sound and that will help us answer these questions:

  • Are 32 speakers enough for fully spatialised sound?
  • Does the hardware work and is it simple to set up
  • What is the experience like building music in this set up?
  • And are slightly thicker broom-sticks sufficient to hold a person suspended in the middle

We aim to have the prototype built for end of July, the work taking place not in Burgundy but in the crypt of a chruch in Holloway, UK, under the auspices of Merijn Royaards. The hope is that he audio hardware will be fit for purpose, and so future iterations will be able to reuse the same hardware.

I’d love to express my thanks to, among other in what was a glorious group project, Tomek Smilok, Alix Faddoul, Iannis Bardakos, and Primavera Di Fillipi for actually building the thing.

Leave a Reply