Homemade distortion device, intended to eventually become part of a synthesizer module.  This is more or less a circuit designed to use up some of my old components - in particular, a pair of antique germanium transistors.  The finished circuit will be a dual distortion unit based on the venerable "Fuzzrite" design modified to work the the components I have and Eurorack signal levels.  What you hear in this recording is two 2N3904 silicon NPN transistors, used for testing because they're cheap.  The eventual final version will incorporate the germanium PNP devices and a pair of silicons as well, but I'm not sure yet what type.

Input signal is the "sine" outputs of two Tiptop Z300 VCOs: first just one at 300Hz, then at about 2 minutes in I add a second at 450Hz (forming a classic "power chord").  These are not very pure sine waves; they have a fair bit of odd-numbered harmonic energy as they come out of the oscillators.  You can hear the input mix for a few seconds at the start and end of the recording.  Everything in between is what comes out of the distorter as I adjust its knobs (and occasionally the input levels).  The Fuzzrite circuit basically consists of two simple inverting transistor amplifiers, each self-biased just a little above ground.  As a result, the lower half the waveform gets clipped in the first stage even at relatively low levels, and then the second stage is driven to clipping almost all the time.

There are knobs for input attenuation (needed at the high Eurorack level - bear in mind the original was for much lower-level guitar signals), output gain (I added an output buffer based on a 1458 dual op amp, another antique), and most importantly, to crossfade between the two stages.  That knob offers a wide range of different distortion sounds as the harmonics from the two stages cancel each other in varying amounts.

The background hum is from power issues in the case I was using at the time I recorded this - not the module's fault, and corrected by a power supply upgrade.

"Shinji" distortion device test

Quick test/demo of my modified Erica Polivoks VCF (kit from https://soundcloud.com/ericasynths ), a recent clone of the VCF from the classic Soviet synthesizer.  It's a state-variable filter built around K140UD12 programmable op amp chips, which are themselves Soviet clones of the American µA776.  Persons looking at the schematic diagram of this filter have been known to ask "What?  Where are all the capacitors?  Is this a joke?"  The trick is that these op amps have slow, and controllable, slew rates.  In this circuit they are adjusted so that their slew rate is especially low, down in the audio range, and that makes them function as integrators to build a state variable filter.  The capacitors are effectively inside the op amps.  It was, like many old Soviet electronic designs, a way to make the most of the available parts in a time and place where hobbyists and even commercial manufacturers couldn't just send off an order to Mouser.  Nowadays, we can do that, but we can't so easily get the authentic chips; I ordered mine from a colleague in Latvia who (as I understand it) made an arrangement with a local semiconductor factory where they still have the original tooling from the old Communist days.

Very basic patch with a clean sawtooth wave going into the VCF, then a VCA; separate ADSR envelopes on the filter cutoff (CV1) and the VCA; sine wave LFO into the filter's CV2 input for most of the track, but that one unplugged around the 2:40 mark to demonstrate one of my mods, see below.

I modded the kit by (among other things) removing all the aluminum electrolytic capacitors from the signal path, replacing the low pass/band pass switch with a potentiometer to smoothly interpolate between the two modes, and normalling the CV inputs to the internal LP and BP signals.  That last was a suggestion from user "SydHeresy" on the Muffwiggler Forum; it means that when there are no cables plugged into those inputs, the knobs for them can be used to add extra chaotic distortion.  You can hear its effect from around 2:40 to 4:40 in the recording, when I unplug the LFO, especially around 3:30.

Also take note of the arpeggio-like things right at the start.  Those are just single long notes, with the filter set on the edge of oscillation and slowly sweeping upward.  That makes it pick out the harmonics in the sawtooth wave one by one, so each note splits into a whole sequence.  Quite an appealing effect, I think, and one I haven't been able to reliably achieve with the other filters I've tried.

Erica Polivoks VCF demo

Start of V/oct calibration testing on the Octothorpe.  The software sends random control voltages to the oscillator, measures the resulting frequencies, and calculates the current response as a number of volts per octave.  Then by physically warming and cooling the exponential converter and watching as the volts/octave ratio changes, I can adjust the temperature compensation circuit to minimize that change.

To make this track, I've run the raw oscillator output (from a single core), as used by the calibration procedure, through an Erica Polivoks filter and a Doepfer bucket brigade delay to make it a little more musical.  This is some idea of how the module might sound in actual use, although its real power, not demonstrated here, comes from the fact that it has eight cores with adjustable detuning between them.

Octothorpe V/oct test

Demo of the Octothorpe's spike and mix outputs, through a modulated Erica Polivoks filter and a Doepfer bucket brigade delay.  These are the kinds of sounds it was designed to produce.

Octothorpe spike/mix demo

Over on Muffwiggler they were doing a "self-patching" challenge - see what kind of music you could make with just one synthesizer module patched to itself.  This is my attempt, with a 4096-stage Doepfer A-118-1.  It's not only a single-module patch, it's also playing all by itself, with no human intervention beyond turning up the feedback at the start and turning it down at the end.

Bucket Brigade Breakdown

Demo of the MSK 006 exponential-response discrete-transistor VCA.  Basic subtractive synthesis patch (two Z3000 oscillators generating slightly detuned sawtooth waves, Dr. Octature II modulated by an ADSR envelope) run through two channels of the VCA in series, one controlled by the MIDI velocity to apply a bit of expression and one to apply the volume envelope.  Multitracked to get the chords.

How *do* you demo a VCA anyway?  This one is meant to be low-fi, so it colours the sound a little bit more than others would, but even so, it's hard to pick out which parts of this track are really the sound of the VCA and which are from other modules.  Still, I hope it at least gives some idea of the kinds of things the module is good for.

If you listen carefully (headphones help) you can hear a relatively high noise floor (probably mostly from the VCAs) and a bit of a "thump" effect, especially on the bass notes, which comes from CV bleedthrough.  There is some harmonic distortion, responsible for a bit of high-frequency sparkle, measurable with test equipment, but it's hard to distinguish that by ear from the high frequency content that made it through the filter.

Plans for building your own will be posted on my Web site eventually.

Flavordynamic Rag (MSK 006 demo version)

Breadboard test of the MSK 007 leapfrog-topology voltage-controlled filter.  This is actually the first time I've gotten it to work as a controllable filter at all.  Still some significant issues with distortion and DC offset, and it's nowhere near meeting the theoretical specs for stopband attenuation, but I think it sounds pretty good even at this stage of development.  Input signals are white noise, 440Hz sawtooth wave, and Ayumi Hamasaki.

Leapfrog Filter (early breadboard test)

Another breadboard test of the Leapfrog Filter.  Since the last one:  fixed some bad connections that were causing distortion; rearranged the op amp integrator circuits to keep LM13700 outputs at virtual ground (improves frequency range and distortion); much debugging and testing. 

In this test, it's first white noise with a filter-cutoff sweep, then a 440Hz sawtooth wave, then those two things again but with resonance turned up.  Toward the end of the second sawtooth sweep the resonance is so high the filter has started to oscillate, and you can hear some zipper-like intermodulation products between the filter oscillation and harmonics from the sawtooth.  Finally, a segment of music (from an old Ray Conniff vinyl record).

Leapfrog Filter (with new op amp integrators)

So.  I wanted to know how the reverb built into Clouds's stock firmware sounds, and although there are piles and piles of demos on the Net, I couldn't actually find one of that specific feature.  There are demos of the Parasites alternate firmware's "Oliverb" mode, and there are "overview" videos for the stock firmware that mention the reverb mode but do not give a clean sample of what it sounds like.

I asked on Muffwiggler whether there was a clean demo of the stock firmware reverb, NOT the Parasites firmware Oliverb mode and NOT just a few seconds in an overview video.  So, of course, the Wigglers helpfully explained to me that Oliverb mode is part of the Parasites firmware, and that I can hear a few seconds of the stock firmware reverb in the overview video.

Anyway, now that I have a Clouds of my own, I've made the demo I wish I could have found.  Here is a sequence repeated through several different settings of the *stock* (not Parasites) firmware reverb mode.  Gaps in between the repetitions so we can listen to the reverb tails.

I have "position" set hard counterclockwise (so there will be no granular delay),  "size" hard clockwise (maximum grain length), and grains triggered by the trigger from the sequence; the idea is to make the granular engine as close to straight-through as possible, so we only hear the reverb.  Blend modes are 100% wet, 50% stereo spatialization, 0% feedback, and trying several settings of reverb, from 0% to 100% in successive segments of the demo.  The sample quality for the granular engine is set to high. 

Clouds reverb test

First recorded test of the MSK 007 Leapfrog VCF prototype.  This is basically the final version of the circuit; the remaining issues to be worked out involve the physical design, knob adjustment ranges, and so on.  For the test I've got a chord sequence controlling V/oct inputs on the Leapfrog and two Z3000 oscillators; the sawtooth waves from those are mixed together and fed into the Leapfrog input, two triangle LFOs feed the linear and exponential FM inputs, and that's it.  A little bit of manual adjustment of the Leapfrog's knobs over the course of the recording to bring out different sounds, but the audio is unedited, just a straight recording of what the Leapfrog does.  Listen for the linear-FM birdsong effect...

Leapfrog sequence test

"Pinging" the Leapfrog filter, with an envelope generator set up to create a very short pulse directly into the audio input, and feedback set almost but not quite to the point of sustained oscillation.

Leapfrog percussion

Words by J. O. Barrett, music by S. W. Foster, 1869.

A "planchette" is a pointing device supposed to be used to communicate with spirits on the Astral Plane.  Nowadays they are usually used with Ouija boards, but this song dates from before the Ouija board was invented, during a time when there was a fad of using planchettes with pencils attached to write out messages (as in the picture).  There were several popular songs written about it, and most of them were humorous, but this somewhat peculiar piece is more like a hymn - somebody evidently took the whole thing VERY seriously!  Simple MIDI piano version, stripped down to one verse and one chorus, just to check that my transcription of the sheet music is right.  I'm working on a more elaborate modular-synth version with my own rearrangement of the music.

Oh, wand of thought, by angels taught
To playful give the fairy story
How bird-like glides on dreamy tides
Life's mystic bark to realms of glory!
Thy compass drifts on music rifts
Across the clear immortal river
'Neath summer skies of rainbow dyes
That lofty span the bright Forever.
Write, write, canny Planchette!
Set the truth-echoes humming;
Write, write, canny Planchette!
Answer angels, coming, coming, angels coming!


Planchette:  Set the Truth-Echoes Humming (MIDI study)

Work in progress - I eventually plan to produce this on a modular synthesizer, but my equipment is on the other side of the Atlantic at the moment, so this version is software synthesis with piano samples.  It uses 22-EDO tuning, which means 22 notes to the octave instead of the more popular 12.

Matt Skala

Popular tracks