Guy de Bievre > Residency Nov 2007 – primitive optocoupling R&D

I’m a composer and musician and I like to use electronics to add unpredictability to my works. For a number of years now I’ve been achieving this using feedbacking microcontrollers (active rather than interactive). Until recently those controllers would communicate with computers and software. But I slightly got fed up with “digital sound” (not with digital control), because regardless of your platform or software all computers sound the same, i.e. the sound of the D/A converters of your sound card (and there isn’t much character difference in them). So I wanted to move to analog sound processing (though I realize with the electronic parts of today this is a relative concept). But how do you control analog interfaces with a microcontroller, without either complex mechanical contraptions or very limited relay circuitry? I went for the easy way out: digital pots…and had to conclude early on that it wouldn’t work as there was no way to avoid signal noise with them. I needed electrical separation between the microcontroller and the analog sound circuitry. One look at my Morley volume pedal (which uses optocoupling to avoid the usual pot noise volume pedals are plagued with) and I knew optocoupling would be the way to go. And STEIM seemed to be the right place for R&D on the subject.

I was glad STEIM seemed to think my work was worth a residency on their premises…The idea was to use a microcontroller (basic stamp, I’m an old fashioned guy, never got beyond linear programming) to control a light source and have that light source “read” by an LDR controlling different stages of the analog circuitry. The first problem that occurred was that on a very simple level the microcontroller cannot multitask, i.e. I could have one light source doing things, but meanwhile couldn’t do anything else: useless. I found the answer to this problem in the 74HC595 (8-bit serial or parallel shift register). And the beauty of the 74HC595 is that you can daisy chain them, meaning that with just 3 of the 16 basic stamp pins I could control n times 8 LEDs. This meant countless combination possibilities, with certain LEDs staying on while others were changing, and meanwhile I could use other pins of the controller to do other things. From there on it was “easy”, the only thing I had to research was the calibration of the LDRs, on both ends: I had to find out the effect of the LEDs (and researched colour combinations) and call upon Ohm’s law to adjust the resistance of the LDRs. After one week I had a satisfying prototype to take home and to turn into a first version, to be used in a new work commissioned by Ensemble Intégrales. In this work, which is for three instruments, two of the instruments go through a separate ring modulator, the pitch of which is controlled by LEDs (7 per ring modulator), another set of LEDs controls the dry/wet balance of each channel (very simple, the LDRs acting as balance pots on the signal input). A third LDR (part of an RC circuit) send its reading back into the microcontroller as a random seed, guaranteeing ever changing, unpredictable LED patterns and determining MIDI CC data that the controller sends to the overall reverb machine.
It all fits snugly in a little box and works like a charm (I tested it during a solo concert I gave at the Experimentelle Musik Festival in Munich early December, and it is now off to Ensemble Intégrales to be used in the premiere of the new work in Hamburg at the end of January).
I do confess that it still isn’t full analog sound processing: there’s the digital overall reverb and the ring modulators aren’t real ring modulators, they’re based on Tim Escobedo’s “Thing Modulator” (cf. ) built around the LMC567 (cmos), and a cmos IC can hardly be said to be analog. Still it doesn’t sound like computer processing and it sounds close enough to a real ring modulator to fool any old modernist. But, now that the concept has proven to be working I’m in the process of building a second version, which will be way more analog as it will control a fully analog wah and a vintage Hammond B3 spring reverb.
You can always contact me if you like to know more about schematics, coding, etc…though everything is fairly straightforward and can easily be found online. (guy(at)
Download a small movie clip here.

Comments are closed.