Trying to recreate Ryoji Ikeda sounds (and failing...)

I visited an Ikeda installation in Berlin in 2017. He’d set up a parabolic horn speaker playing simple sine waves tuned to the room, so if you walked around - or just turned your head - the sound rippled, as the reflections interfered and reinforced the direct sound.

A big aspect of his work is that kind of manipulation of space with sound. Sometimes it’s with an installation, sometimes it’s with things meant to be heard on headphones. The music on Dataplex sounds really different on speakers and headphones.

The biggest thing I hear on “Dataplex” that you might recreate in Rack is this: instead of using VCAs for notes he’s using square waves as envelopes. That ‘instant on’ aspect means there’s voltage discontinuities that manifest as clicking. You could try this by multiplying square waves with audio signals.

@nebdaar that’s interesting, I’ll try it. Thank you for the suggestion.

@chaircrusher thank you for the suggestion on square waves as envelopes, seems reasonable! I’d love to see / hear an Ikeda installation / concert in person…

Here’s what I came up with in a few minutes:

2021-02-18.vcv (380.9 KB)

http://cornwarning.com/chaircrusher/Chaircrusher-FakeIkeda.mp3

This does the thing with using pulses as envelopes and multipliers instead of VCAs. I don’t think it’s as good as Ikeda, but it does sound like him. The pulses driving the noise and two sine voices come directly from the Clocked device, and you can tune how long they last with the pulse width of the clock.

I use flip flops to get longer note durations out of the trigger stream from the rightmost SEQ3.

It’s not going to be a Top 40 hit, but it illustrates using very primitive logic to generate blips and tones.

I brought it into Sound Forge to edit the ‘top & tail’ and the waveform of the track is BRUTAL looking. I didn’t do anything to the output of Rack on this one. It’s exactly what it generated in real time.

1 Like

I wonder listening to Ikeda’s stuff whether he’s building simple stuff with logic gates and timers as tone generators. Or doing something similarly primitive (in the audio domain, not the conceptual domain) with Max MSP.

Reminds me of a guy who played live and had a long piece that was all sine waves, but he sampled them so that every note had a discontinuity click at the start. He made that happen by sampling a sine wave and cutting it so it didn’t start at 0 volts.

1 Like

The Left/Right assortment is a bit more complex than that. As described: each group of sounds has its own set of rules and placement.

Take Data.Syntax There is a high pitched sine running constantly at the beginning It is 70% right, 30% left ; there is slightly different EQing (left side: mids are dimmed) The sine is sidechained to bass - and ducks it 100%

The ticking is indeed made by sending a very short pulse to an Amps volume. It is slightly leaning to the right and branches into 3 general variations:

  1. the original signal: set slightly right and at ~70% velocity/volume ; pan does not change
  2. panned hits at higher velociy. right side is generally louder and panned closer to center. they vary slightly in velocity settings. left and right side are differently EQed… with the right more saturated.
  3. irregular variations. these are varying in panning, velocity and length (longer pulses sent to amp.) Fin: Towards the end a saturated full velocity tick runs center - replacing the variations.

There is a “dwuuh” sound that jumps through panning positions (prioritising the right side, as all sounds here). The ticking is set to duck the sound (muted). The ticking towards the end does not duck - in response the “dwuuh” starts fading out and reduces panning.

The Bass has fixed positions left and right left: a desaturated variant also with harder panning. right: a saturated variant. farther to the right than right sides ticks, closer to center than left side bass panning, velocity and sound are fixed center: bass has overall less velocity and saturation than left and right - in exchange plays varying pitches (MIDI Notes/v/Oct).

In general: Left and Right are not Delayed Repeats but distinct sounds and/or signal paths. Louder, longer Sounds tend to have fixed positions with fixed velocity by rule of thumb - that means if they have more than 1 position i.e. left, each position is coupled to set a set velocity, EQ setting and other parameters. There can be 4 variations left - but each will always appear at its respective panning. [Bass, Synth, Stabs…] This establishes a sense of space in two directions.

The positioning = bound to settings holds true for most sounds. But varying velocity is instrumental for these [mostly percussion and fx]. By reducing or raising velocity in recurring manner at fixed pannings for distinct sounds… the sense of space extends up- and downwards.

Now Panning over time is introduced to some sounds. They begin to travel (most often slow and slightly it seems).


I’m not too knowledgeable about sound design but: You achieve this effect by

  • Planning for 3 Signal paths for every Sound [Except Pads]
  • Signals that always switch left/right should go through a switch (i.e. Bogaudio switch)… use AS Steps to “count” and flip the switch.
  • Use RCM Duck to easily sidechain Sounds
  • VCV Mid/Side is a necessity for processing all Pads and Synths and Bass
  • Vult Decline & Tohe are light-weight options for EQing
  • FrozenWastelands Compressor is a must (it can do Mid/Side compression and takes CV for everything)
  • Submarine PG-104 and PG-112 are good to Pulse Generators… use them with Vult Punch (Pulse to Punch CV in, Sound source in, Punched out sound out) to get your sounds.
  • Use MindMeld EQ & Groups to easily apply general variance for i.e. “all percussive sounds left” and “all synth [foreground] sounds left”.

Panning: if you don’t want to use 3 channels for each sound you can, after giving each sound its own blend, merge them. and use SV Multipan to pan them. Output L and R into seperate VCV VCAs and use BogAudio PolyCon or PolyOff to control each signals velocity… from there into the Mixer

PRO TIP: Use Rackwindows Console. Even if you have to split your signal and manually connect each channel. It will add space between the sounds without adding color, distortion or saturation. you’re trying to close in on a sound that is carefully crafted with professional equipment and expertise. this’ll help.

Thank you a lot, what a great analysis!

I’ll have to read it again hands on with vcv open, but I think it will help a lot. Thank you!

I’m not sure to understand why to use the rackwindows console. What do you mean with “it will add space between the sounds”?

Rackwindows Console makes use of the Console 6 or True Console (can set in context menu) Algorithms. Console is part of a bigger set of VST Mixing Plugins.

In VCV summing mixers just add the inputs together. The result is often rather flat. Console applies its Algorithm on each input before summing, setting them apart. The more inputs are added, the stronger the effect. Very similar to an analog mixer.

Wow thank everybody for the ideas!

Oh wow, thank you!

http://www.raisedasapackofwolves.de/2021-03-30%2001-22-46.mp4

here’s a Ikeda-Style Patch. Recording is laggy unfortunately.

No news on the topic itself. Fitting nevertheless I think.