Modular analogue synthesis as an extension of a digital system, part 1

I’ve been deep into SuperCollider for a long time now – 14 years and counting. SuperCollider is a cross-platform text-based software environment with its main focus on sound synthesis, which also contains a vast toolkit for audio analysis and algorithmic composition. It’s powerful and efficient: a quick test shows me I can run 5000 individual sine oscillators on my average spec desktop (2.5 ghz i5.) The help system shows over 1200 UGens available (“unit generators”, i.e. oscillators, filters, triggers or other generators or processors.) Together with a comprehensive GUI system, there is little it can’t do.

However, I’ve always had an eye on modular synths, mainly for one reason: simple analogue filters (low pass, high pass, etc.) just “sound better” than digital ones of the same specification.

That’s not to say digital filters can’t get close, or even fool the ear, but in general when doing the traditional style of subtractive synthesis, there is something about the sound which leaves digital synths sounding a bit “flat.” Now, modular synths can’t play 5000 oscillators at once without requiring several rooms of equipment, or do FFT synthesis, or model natural processes to the same degree of detail, or any of the other things which an environment like SC excels at. But provided they are not over-engineered (some latter day analogue synths contain oscillators that are so stable, they may as well be digitally derived) they can sound really lively in a way which is difficult to achieve inside a computer.

So, how can this analogue synthesis system be controlled? For someone with a lot of experience of digital methods, this area can be seen as a massive drawback. It’s possible to use a MIDI to CV converter – taking the early 80′s technology able to represent all of 128 semitones or 128 steps of a parameter control and convert it into the voltages required to play a tone of a certain pitch or timbre. That’s fine for certain kinds of music, many of which I’m fond of, but in a digital system there is a conceptually infinite world between each semitone.

Recent discoveries have shown that it’s possible to use certain computer soundcards to generate the kind of control voltages usable by analogue synths. This removes the need to funnel musical concepts through the constraints of MIDI. Together with a flexible system like SuperCollider, one could have the best of both worlds: very precise control of the apealling and slightly unpredictable sound of analogue synths.

Having recently taken posession of such a soundcard and a small rack of analogue modular synthesis modules, mostly filters, I have been experimenting with these concepts. Here is a summary of my initial aims:

• to create analogue drum sounds

• put a digital oscillator through an analogue filter in a more-or-less traditional version of subtractive synthesis.

• calibrate the filters and use them as oscillators, including frequency or other modulations.

• take my simplistic but great sounding speech synthesis system based on one low pass and two band pass filters, and perform the synthesis on the modular.

• take my similarly simplistic but great sounding analysis / resynthesis system and resynthesize sounds on the modular.

• freely draw shapes, put these into buffers and control the synth modules in a kind of “oramics” style.

• make a vocoder, with audio analysis taking place inside the computer, and synthesis on the modular.

I’ve already tried most of these things, and will write them up one by one in the coming days.

This entry was posted in code, music and tagged , , , . Bookmark the permalink.

Leave a Reply