Understanding Ghost Electronics
(In Search Of My Muse #1.07)
Historically, the concept of "ghost electronics" was pioneered by the famous tandem Morton Subotnick/ Don Buchla in the late 60's: it was - and still is - an ingenious way to store huge amounts of control voltages on tape (or in digital form).
Back then, the only choice available to store much needed data was on a tape medium. Alas, the frequency response of most tape recorders was not adequate for recording DC control signals. Therefore, a special AM technique had to be used to record these control voltages on tape. For that task, an audio frequency (carrier) had to be modulated in amplitude by the control voltage to be encoded (the resulting modulated signal was then recorded and demodulated upon playback).
Two analog modules were the key to a successful CV/Gate encoding/decoding procedure: the envelope follower -also called envelope detector - and the pitch-to-voltage converter. As a reminder, an envelope follower is a device which produces a voltage proportional to the instantaneous amplitude (global envelope) of an audio input signal. This voltage is then used, for example, to control the cut-off frequency of a VCF and/or the gain of a VCA, in place of the usual envelope generator. A pitch-to-voltage-converter is often used in conjunction with an envelope follower: this device converts external audio pitches-from microphones, instruments pick-ups or tape recorders- to DC voltages. These can then be used to control the frequency of a VCO and/or a VCF (1Volt/oct.inputs) or any other module in your patch.
In the above example, a 500 Hz. VCO sine waveform (carrier) is patched to the audio input of a VCA. The output of a fast clock is used to trigger an envelope generator, having a sharp 'Attack' and a short 'Release' (percussive envelope). The output of the envelope generator is then patched to the gain input of the VCA, whose output is connected to a mixer. The combined signals (carrier + modulating signal) are then recorded on Track 1 of a multi-track tape recorder (see example a).
After hearing the results of the track you just recorded, you decide that this train of raw pulses is too long and should be adapted in a more dynamic way.
Connect Track 1 output to the audio input of a VCA. Then, patch a keyboard gate output to an envelope generator, whose output controls the gain of the VCA. Set Track 1 in 'play' mode and experiment. Now, every time you'll depress a keyboard key a 'sample' of the original raw train of pulses will be taken: you now have individual "chunks" of train pulses that you can adapt for your gating needs. Record the results on Track 8 (see example b).
Later on, you decide to mix the original Track 1 output with the new Track 8 you just recorded and PAN, in real time, between the Left and Right channels. Record the results of your 'live performance' on channel 6 and 7, respectively (see example c).
When done, listen to the tracks you just recorded. In case of timing errors, record "blanks" to remove the erroneous segments or filter the unwanted signals with a sharp comb filter.
The next stage is the actual decoding of the signals on tape. For clarity sake, let's assume you have only recorded one track of control signals (see example d).
In this case, the tape recorder's track 1 output is fed into a sharp band-pass Filter (VCF), whose cut-off frequency is set to 500 Hz (to match the frequency of the original carrier signal). The output of the BP Filter is patched into a pre-amplifier, whose output is fed into a limiter (a device with Automatic-Gain-Control or, eventually, a manual attenuator). The output of the limiter is then patched to the input of an envelope follower, whose output is fed into a low-pass VCF set for a 500 Hz. cut-off frequency (to smooth the final curve of the control signal). Finally, the resulting control voltages are then sent to various devices for further processing,
Note: the above example is only valid for one track of decoding: If you have recorded CV/Gate signals on all 8 Tracks of your tape recorder, you will need a lot of analog gear to decode them in real time... Indeed, mixing all encoding signals together and decoding once after mixing will NOT give satisfactory results: the envelope follower will only detect the 'global envelope' representing the sum of all the recorded channels coming from your multi-track tape deck!
"Voicing" Your Pulses
Needless to say, this AM encoding/decoding technique is not strictly limited to the analog domain. If you own an 8 track digital recorder, you can use this technique to store/retrieve control voltages or gates to-or-from a CD-R track (the digital signals should be first fed-back to the analog world through a Digital-to-Analog converter and then decoded as shown).
Evidently, you can also use your voice - or any other signal generated by an acoustical instrument - to "phrase " your CV's or Gating needs in a creative way (see middle illustration, in example a).
Note: Joan La Barbara (Morton Subotnick's wife) used 'multi-phonics' singing techniques to generate vocal pulses for many of her husband's "ghost electronics" compositions.
For example, let's assume that you want your voice to control the L/R panning of a sound, with a 90° out-of-phase train of pulses. To do that, use two microphones set at a 90° angle and generate a " TTTrrrrrRRRRRttttt " glottis sound, while moving your head back-and-forth between the two microphones (phase differences). When done, decode the resulting pulses and use them in your patch to Pan/Cross-fade your complex sounds in a stereo or quadraphonic field.
(Excerpts from my tutorial 'The Art Of Analog Modular Synthesis by Voltage Control).
André Stordeur has taught analog modular synthesis since 1973. He studied with David Wessel at the I.R.C.A.M, in Paris, and with American composer Morton Subotnick.