Brian Eno released Discreet Music in 1975; the majority of the album is taken up by its 30-minute title track, which was one of Eno’s first experiments in the quiet, unobtrusive music that would go on to define his career, as well as ambient music. The songs concept was conceived by Eno when he “discovered” a new way of listening to music, where at low volumes the sound is on the verge of vanishing, and merges with the background noises of the environment.
In this article I’ll explore how Eno created Discreet Music using a simple but clever system that utilised early synthesizer sequencing and sound-on-sound tape looping. I’ll also recreate the sounds using Arturia Synthi V and Soundtoys Echoboy.
Recording Discreet Music
Eno recorded Discreet Music using an EMS Synthi AKS, a portable analog monosynth released in 1972 with modular and sequencing capabilities. Eno was a big fan of the EMS synths; he famously used an EMS VCS3 on David Bowie’s Heroes, and they were his synths of choice until the Yamaha DX7 came along in the early 80s.
As well as the digital sequencer, which I’ll discuss more in-depth later in the article, the EMS was also notable for its unique patching system. Whilst many modular synths of the 70s used large form factors and patch cables to make complex routing possible, the Synthi used a patch matrix where the user would place pins to create connections. Different coloured pins could be used to control the amount of attenuation and modulation. This setup made the synth small but powerful. Other interesting features are the joystick, allowing for hands-on controls of certain parameters, and the built-in spring reverb.
Arturia recently released a software version of the EMS Synthi, called Synthi V, which recreates the EMS Synthi, with some welcome modern additions. The Discreet Music patch is relatively simple to patch, and makes use of oscillator blending and the onboard reverb. Start by initialising Synthi V by opening the menu and choosing New Preset… to start a fresh patch.
For oscillators, the sound is a mellow sine/triangle wave made more interesting by mixing a small amount of sawtooth oscillator. On the Synthi, each oscillator outputs two different waveforms, which are routed to the filter via the patchbay grid. Oscillator 1’s sawtooth is patched by default, and we can also patch oscillator 1’s sine wave and oscillator 2’s triangle wave by placing a pin on the matrix dots 3H and 6H, lining the oscillator sources on the left with the filter input above. Balance the volume of the sawtooth wave by lowering oscillator 1’s saw level knob down to the 3 mark.
Set the filter by lowering the blue frequency knob in the filter oscillator section to just below halfway (0.472). I also added some subtle vibrato by patching oscillator 3’s triangle to the oscillator 1 & 2’s control inputs (I8 and J8), setting oscilator 3’s frequency nice and low, and the triangle’s volume level low to control the amount of vibrato modulation. Eno also would have likely added some built-in reverb from the Synthi, which will add ambience as well as darkening the sound, so raise the reverberation mix knob to 9.
- Discreet Music Synthi Dry 00:00
I made Discreet Music using (by today’s standards) the primitive tools of early electronic music: an AMS synthesizer, with a simple sequencer, a graphic equaliser (this allows you to modify the timbre of the synthesizer’s output), a Gibson Echoplex and 2 Revox tape recorders. All of those tools were quite fallible.
The synthesizer tended to slide gradually out of tune as it warmed up and anyway the keyboard was not a standard ‘equal temperament’ keyboard such as you’d find on any commercial synthesizer now: you began work with it by tuning the octaves, which, along with the sliding pitch, left some leeway for variation.
The Graphic Equaliser was slightly the worse for wear and some of its faders crackled, so I didn’t use those frequency bands during the recording of the piece. The Echoplex used a loop of magnetic tape which rather erratically circled round a series of playback heads. The tape was old and didn’t reproduce high frequencies; the ‘wow and flutter’ produced by the slippage of the tape created a gentle and sporadic chorusing effect on the echoes it delivered. – Brian Eno
In Eno’s Discreet Music system, the EMS synth is run into an echo unit, a graphic EQ, and finally the dual tape system. The echo unit was a Maestro Echoplex, a classic tape delay, used to add more ambience to the synth sound. Soundtoys Echoboy has an Echoplex style that sounds great. Set the delay time to 160ms, with a high feedback and mix level of 40%.
The next stage of Eno’s system is a graphic EQ, which was his sound manipulation input during the Discreet Music recording process. By raising and lowering certain bands, Eno could subtly affect the tone and timbre of the synthesizer before it reached the loop phase. Note that when you make changes in this way, the results aren’t immediately heard, as the EQ was placed before the looping system in the effects chain.
Next the signal gets to the dual tape machines, or Frippertronics. Although the idea of using two tape systems running into each other may seem complex, it was just a way of creating longer delay effects than the hardware of the time allowed. Eno’s dual tape system allowed up to several seconds of delay, similar to a modern-day loop pedal with the loops gradually fading out.
We can use another instance of Echoboy for the Frippertronics, this time using a long delay time of 5650ms, with mix set to just below 50% and feedback set high. Now you can hear that each phrase repeats after a five-second delay, causing a cascading effect and adding an extra layer of depth to the piece.
- Discreet Music Frippertronics 00:00
Much of what makes Discreet Music interesting is that the music is made up of two independently running sequences. The two sequences are panned separately left and right, and loop at slightly different lengths. The two loops interact to form different combinations upon each repeat. This is a technique I discussed in much more depth in my Deconstructing Brian Eno’s Music for Airports article, where I showed how the pieces 2/1 and 1/2 used at least 8 independently running loops to create the illusion of a more complex piece. Discreet Music, being a predecessor to Music for Airports by three years, is an early experiment of this same method.
In Discreet Music, the left channel sequence is approximately 63.5 seconds long, and the right channel sequence is approx. 68.7 seconds long. The 5 second difference between sequence lengths means that upon every repeat, the beginning of the right channel sequence lags another 5 seconds behind the left channel sequence. It takes roughly 15 minutes of playback for the two sequences to roughly sync back to their original alignment.
The Synthi’s onboard sequencer is one of the earliest digital sequencers, powerful for the time but unusual by today’s standards. You can record a sequence by simply pressing the record button and playing in notes. Rather than being a traditional step-sequencer however, the sequence length knob sets the time taken to play the sequence, which is done by adjusting the speed at which the sequence plays back.
Higher speeds mean slower playback, and shorter speeds mean faster playback (think Pink Floyd’s On The Run). To sequence Discreet Music, Eno would have set the sequence length to the longest time and played in the sequence melodies manually.
Below you can watch two instances of Arturia Synthi V playing the Discreet Music sequences. Each track has been processed with Soundtoys Echoboy, for standard delay and for Frippertronics-style looping.
Half Speed/Double Speed
”I have a theory that, as a maker you tend to put in twice as much as you need as a listener. It’s the symptom of contemporary production. That’s why old records are interesting, because they don’t have that problem a lot of the time. With the facilities that you have today, you tend to plug every hole… you’re always looking for that charge, so you put more and more in to get it.
But as a listener you’re much less demanding… you can take things that are much simpler, much more open, and much slower. It’s often happened that I’ve made a piece and ended up slowing it down by as much as half. Discreet Music is an example: that’s half the speed at which it was recorded.” – Brian Eno
In this interview, Eno reveals that Discreet Music was slowed down by half after the recording process. This would have been a necessity, as the Synthi V’s onboard digital sequencer was only capable of sequence with a maximum length of roughly 34 seconds. When lowering the speed of tape by half, you also get a reduction in pitch by an octave, meaning that Discreet Music was also recorded an octave higher than we’ve been working so far. There are also several other side effects of slowing down tape recordings, such as loss of high-end frequencies, and unstable pitch and speed.
Below is an approximation of how Discreet Music sounded before being slowed down. The main changes are the oscillators are tuned an octave higher, the Frippertronics delay time is 2825ms, and I also opened the filters slightly as the original sound would have been brighter.
Thanks for reading! In the download button below you can find the Discreet Music MIDI sequences, as well as the Synthi V patches and Echoboy settings. Experiment and enjoy!
17 thoughts on “How Brian Eno Created <em>Discreet Music</em>”
This is a very cool and informative article. It shed some light on a few questions i had. Wish more people took time to analyse these sorts of musical mysteries
So, there’s one thing you didn’t address in this article— Eno only had one Synthi. I guess we must assume the two 30 second sequences were recorded separately and then mixed together. But what I’m wondering is, were they recorded dry on two tracks and then treated with Echoplex and Time Lag together? Or were they each recorded wet with Echoplex and Time Lag, and then the two “wet” recordings mixed together? Or some combination of recording and processing and then mixing (and processing)?
I think recorded wet with Echoplex and Time Lag and then mixed together. It’s the easiest way of doing it (the mix is stereo but the Echoplex only takes mono ins and outs) and the most creative, as he’d be able to monitor and hear the Time Lag effect in real time.
Thanks for this fine article and the midi downloads.
Are there wet versions of the two Synthi sequences available somewhere?
Amazing and exact replication. You’ve got all the bases covered. Something so complex, so evocative, created with simple loops and delays and tape.
I was struck by the shelf in the photograph of Brian Eno. Are they textbooks, or tape boxes? That’s a lot of tape boxes.
reel-to-reel tape boxes, I imagine he went through a fair amount of tape
Great to put a light on the mysterious piece ! would be also interesting to focus on « thursday afternoon », thanks for all of this !
Great. Thanks a lot.
And another wonderful insight into the systems of Ambient music. Thanks alot for your work – again it’s very inspiring for me to start experimenting with this knowledge. Had a lot of fun in my studio with it yesterday. I am getting more and more into generative music. Thanks to you! Next I want to integrate my hardware synths more and make use of the tape function from my OP-1.
Very cool! Makes me want to get my hands on that Arturia Synthi too.
Brilliant decoding and recreation. This is a great Window into the power of limitations. Love it. Thank you !
excellent and clear review, love Eno and have often wanted an in-depth analysis of this simple yet powerful composition
Thanks Paul! Yes, I’ve always been curious about the exact sequences too, turns out it’s simpler than it sounds!
excellent review !