Tag Archives: music

Jam of the day: Not your dad’s vocoder

The Mutable Instruments Clouds is one of the most popular Eurorack modules ever. It was intended to be a “texture synthesiser”, an instrument which uses granular synthesis in order to extract a texture from an incoming signal.

Granular synthesis is just the “default” mode, however. There’s other ways of extracting texture from sounds and sure enough, buried on secret firmware number 4 is a happy little phase vocoder! Rather than explaining how it does what it does, take my word for it that it can smear any dynamic sound into a static timbre using complicated maths which gives engineering students lifelong PTSD. I was using this trick to create drones back in 1999 by “misusing” the noise removal feature of Cool Edit Pro.

Anyway – sound goes in, drones or whooshes or wibbly noises or crackles come out. Here’s what it sounds like.

There is no actual synthesiser in this recording aside from Clouds – I played a D5 chord in on the electric mandola (2200 if anyone’s keeping score) and froze the timbre in place to create the drone. Then I played in some other notes live over the top. Add a bit of delay to smooth it out and that’s all there is to it!

Despite its popularity, the designer of Clouds discontinued it earlier this month, citing disappointment that “feedback and reverb got stuck to the maximum setting, resulting in a never-decaying smudgy howl of hype”. This is part of the reason I avoided it for so long, because I didn’t want to.

Patch of the day: Wire speaks to wire

Reverb is an electronic effect which gives any sound signal the illusion of reverberation. Spring reverb does it by electromechanically vibrating one end of a set of springs in a tank, then recording the vibrating strings to create a pleasantly “wet” metallic ringing.

Behold my happy little Belton spring reverb tank, ready to be patted soothingly.

This short ambient piece was made using a spring reverb effect, but it’s not actually reverberating anything except itself. I sent its output back into the input to create a feedback loop. This made the springs ring uncontrollably after a few seconds, but I can mute the springs with my fingers to quieten them down.

There’s been a bit of reverb and delay added afterwards to space it up a bit. 🙂

Jam of the day: Brainburner (excerpt)


This is a jam for anyone who enjoyed the harder house tracks like Rollin’ and Scratchin’ or Rock’n Roll on Daft Punk’s album Homework. It’s a spiky glitchy drone with some nice polyrhythmic percussion and that’s it – maybe not everyone’s cup of tea but I like it. 🙂

Tech notes

I used two of my new modules for this! The Music Thing Modular Pulses mk 2 is outputting a randomly generated six note rhythm tapped at two points for the clap and the hi-hat. The modulation state of the drone is being sample-and-holded in sync with a Euclidean rhythm out of the Little Nerd. The drone is feeding into the Voltage Control Lab VCF-74 Mk 1 where it’s being frequency modulated; that’s fed into a Ladik Waveform Animator then on to the Eowave Fluctuations Magnetiques filter bank through a parallel low-pass and band-pass filter.

The drone is just sitting at one steady note. So why does it sound so lively? The experiment was to see what would happen if I put a pulse wave through a low pass filter and then put that through the waveform animator. The animator only works on waves with gradual transitions which means a pulse wave goes in and out unchanged, but a low pass filter can introduce those transitions into the pulse wave by filtering the sharper frequencies away to make it transition more like a sine wave. The texture of the drone comes from crazy chaotic interactions between the filter frequency modulation, the waveform animator and the pulsewidth of the original oscillator.

Patch of the day: They interrupt ladies and gentlemen

Have some sound art.

The original sample is a snatch of the Mercury Theater’s infamous “War of the Worlds” broadcast, looped to twist its meaning. The chief glitch mangler effect here is a PT2399 digital delay chip living inside the Befaco Crush Delay v2. Its delay length is being automatically modulated by four stacked modulation sources – two random, two cyclic – via a Befaco A*B+C attenuverter/offsetter. The strength & offset of the delay modulation and feedback strength of the delay were manually manipulated during the recording.

tl;dr = it’s like autechre but funnier

Sample source from http://mercurytheatre.info/

Introducing DASYRAC!

What’s a DASYRAC then?

DASYRAC is my voltage-controlled Eurorack modular synthesiser. It’s a synthesiser because it creates positive and negative voltages which turn into vaguely musical positive and negative sound waves when they’re sent through a speaker coil. It’s modular because it’s built up from modules and instruments which each have a specific role to perform in the process of synthesis. It’s Eurorack because the modules use Doepfer’s Eurorack specfications for size (3U) and voltage (+/- 12V & +5V supply rails, plus gate and pitch CV buses which only Doepfer’s modules seem to use).

Finally, it’s voltage-controlled because the modules are connected by patch cable and signal one another using voltages. Signal types include:

  • triggers – a very short pip of steady positive voltage used trigger a modulation source or function generator, reset a sequence, etc
  • clocks – a series of short triggers used for timing, usually evenly spaced
  • gates – a sustained steady positive voltage which is held “high” for the length of a note or other event
  • modulation – a non-steady voltage which is used to control module functions
  • pitch – a voltage which is used to represent the pitch of a note, where one volt corresponds to one octave (1v/oc); usually steady, but not always
  • audio – a varying voltage which rapidly swings positive and negative and is intended to be heard at some point

Because they’re all made of the same stuff (flowing electrons), these signals can play one another’s roles – audio and gate signals can both be a modulation source, sufficiently fast clocks can be an audio source, audio can also be a clock source, pitch can be a modulation source, modulation can be a pitch source, and triggers can even be an audio source as long as you’re into snaps and clicks. Part of the fun of modular synthesisers is just seeing what happens when you connect modules together in unconventional ways.

Here’s a recent photo of a cable-free DASYRAC with colour-coding to show what the different modules do. If a module has more than one function, it gets more than one colour.

A rainbow of possibilities!

The multicoloured thing at the bottom left is a Make Noise 0-Coast synthesiser – it’s not a Eurorack module itself, but it can patch into and out of DASYRAC just fine. As you can see, it’s like a tiny self-contained modular system in and of itself. There’s also an ARP Odyssey floating around somewhere.

Everything starts with a supply of tasty electrons from two power supplies (in red). All the active modules connect to the two power supplies at the back by ribbon cable.

  • Middle row: 4ms Row Power 30
  • Bottom row: Tiptop uZeus

Maybe I’m being too specific about what makes noises…

The green modules are audio sources. They speficially create oscillating voltages which are intended to be heard aloud at some point. Some are more fully featured than others – the green module in the centre top row is an entire self-contained instrument – it’s a recreation of the clap sound from a Roland TR-909. Audio sources which produce a musical tone – oscillators – have a 1V/Oct socket which takes a pitch voltage. The green module at the top left is a sound playback instrument.

  • Top row: Music Thing Modular Radio Music (digital sample player); Music Thing Modular Turing Machine (noise source); Mutant Clap (TR-909 clap instrument)
  • Middle row: Befaco Even VCO (saw-core analogue oscillator)
  • Bottom row: Doepfer A-110-4 QZVCO (through-zero quadrature sine wave oscillator); Befaco Crush Delay (crackly lo-fi digital noise); Bastl Instruments Noise Squared (digital square oscillators and noise sources)
  • Tabletop: Make Noise 0-Coast (triangle and square oscillator)
  • Not shown: Korg ARP Odyssey (fixed-architecture synthesiser which can be run as a plain CV-controlled dual oscillator if need be)

The blue modules are “non-audio” voltage sources and include function generators like the Rampage, step sequencers like the SQ-1, procedural sequencers and clock sources like Tuesday, and random voltage generators like Turing Machine. Their job is to provide discrete or continuous voltages which can drive melodies, rhythms, filters, or whatever else accepts voltage control. If they’re running fast enough, voltage sources can also be used as oscillators in their own right.

  • Top row: Befaco Joystick (gate source and fun voltage generator); Music Thing Modular Turing Machine (random voltage & trigger generator); Hexinverter Galilean Moons (two slope/attack-sustain-release generators)
  • Middle row: Befaco Rampage (dual complex function generator); TINRS Tuesday (procedural phrase sequencer and clock source)
  • Bottom row: Sonic Potions Mal-2 (random smooth voltage generator)
  • Tabletop: Make Noise 0-Coast (MIDI-to-CV converter; low frequency oscillator source; clock generator; stepped random voltage generator; simple slope function generator; attack-decay-sustain-release envelope generator); Korg SQ-1 (simple sequencer)
  • Not shown: Arturia Beatstep Pro (complex sequencer with 2 x gate, modulation and pitch channels, 16 x trigger channels, clock source, etc); Arturia Keystep (keyboard with gate, modulation and pitch out ); Korg ARP Odyssey (gate, trigger and pitch source)

Many of these blue things make excellent audio sources in their own right..

If the voltage sources put out voltages which aren’t quite right, that’s where the orange modules come into play. These modules are for manipulating voltages in different ways.

  • Top row: Befaco Joystick (adds/subtracts incoming voltage); Befaco Dual Attenuverter (multiplies incoming voltage then offsets it)
  • Middle row: Worng Electronics 3×3 mult (routes incoming voltage to multiple places); Thonk AT-AT-AT (triple signal attenuator to make things quieter); Befaco Rampage (turns gates into attack-sustain-release slopes); Erica Synths Pico Input (dual signal amplifier to make things louder); Mutable Instruments Links (makes exact copy of incoming voltage and sends to multiple destinations); Doepfer A-151 Sequence Switch (voltage controlled signal router); Befaco Slew Limiter (limits the rate of change of incoming voltage)
  • Bottom row: CMF Half-Wave Rectifier (splits off positive and negative halves of incoming voltage); Befaco Sampling Modulator (samples and holds incoming voltage when triggered); Sonic Potions Penrose (quantises incoming voltage to equally tempered steps); Worng Electronics 3×3 mult
  • Tabletop: Make Noise 0-Coast (signal attenuverter and summer)

The yellow modules are the effects and filters. Filters in particular make the incoming sound quieter at particular frequencies and also colour the sound in characteristic and desirable ways. They are “voltage controlled filters” (VCFs) because they take an input voltage to control their cutoff frequency, which is the frequency where the filters take effect. There’s also some effects modules for even more retro colour. Many of these modules can be made to create noises of their own without any input,  which also makes them audio sources in their own right!

  • Top row: RYO Aperture LPG (low pass gate, also usable as a VCA); Music Thing Modular Simple EQ (two-band EQ and EQ “tilter”)
  • Middle row: Befaco Chopping Kinky (dual wavefolder); Doepfer A-106-6 XP VCF (state variable voltage controlled filter based on the Oberheim Xpander); Doepfer A-199 Spring Reverb (mechanical reverberation effect)
  • Bottom row: Vintage Synth Lab VCF-74 (serial highpass and lowpass filter based on the Korg Minkorg 700s); Befaco Crush Delay (circuit-bent delay effect)
  • Tabletop: Make Noise 0-Coast (integrated wavefolder and waveshaper)
  • Not shown: Korg ARP Odyssey (switchable vintage low pass filter with CV cutoff control via the pedal input)

Violet modules are a special thing called VCAs – voltage controlled amplifiers/attenuators. A VCA takes two signals – an input and a modulating signal – and outputs a quieter/louder version of the input signal with amplitude/volume tracking the modulating signal. VCAs are where a long steady beeeeep sound is contoured into a short pip or a long mournful wail. Even better: a quadrature VCA can invert the incoming signal – chuck a sine wave into it as a modulator and you’ve got Dalek voices.

  • Top row: RYO Aperture LPG (also usable as a VCF); Hexinverter Galilean Moons (with integrated envelope generator)
  • Bottom row: Befaco A*B+C (dual full quadrature VCA with offset, also usable as a mixer); Bastl Instruments Skis (character VCA with integrated envelope generator)
  • Tabletop: Make Noise 0-Coast (VCA stage including overdrive)

You can never have too many VCAs..

The last set of modules in indigo is for combining signals together. The modules can mix audio as well as control voltages.

  • Top row: Bastl Instruments ABC (dual three channel mixer); Worng Electronics LRMSMSLR (left-right to mid-side encoder/decoder)
  • Middle row: Doepfer A-138a linear mixer (four channel mixer with CV source on channel 1); Mutable Instruments Links (two channel unity gain mixer; three channel mixer)
  • Bottom row: Doepfer A-185-2 Precision Adder (for combining pitch signals together, e.g. to transpose them)

Overall, DASYRAC is an eclectic system with features from the mainstream East Coast school (character filters, rich saw/pulse oscillators), the artier West Coast school (waveshapers, function generators, sequencers, low pass gates), old timey classics (spring reverb, shift register noise) and fun new possibilities (digital audio playback).

If you want to hear the sounds I’ve been making with it lately, check out my SoundCloud.

In these rough storyboards, Pointy has just asked Gronky a pretty curly question – and with that, here’s my activity report for the week 3 to 9 November. It’s on the long side but only because I’ve been busy. 🙂

In summary: the rewrite is going well, I’m working on a C64 version of the music for the end credits (mockup here), and I am still a nerd for the C64’s sound chip some 29 years later.

Prior to last week at the end of October I’d started the process of rewriting – that ramped up bigtime this week to the point where I was safe to start storyboarding on Thursday evening. As of now, 7:30pm on Sunday, I’ve fully boarded 3 of the 5 sequences as currently planned (intro, payoff and outro) and mostly boarded another (lead-up to the gag-fest).

There’s one sequence that remains completely unboarded – that would be the gag-fest, which is what everything builds up to. Right now it’s still stuffed with first-draft ideas. The ideas are mostly funny but I’m quietly prepared to dump any of them if I think of funnier stuff or something doesn’t work in context. I’m aiming for a steady escalation of looniness.

Also I have something like a satisfying ending now. It’s weird to admit this but I didn’t have anything like an ending before. The story just sort of ended – too early, as it turns out. Following Gronky and Pointy for just a couple of extra beats has made a huge difference.

By far the biggest change in the boards from the last version – not immediately obvious from this post’s example images – is that I’m fearlessly using super-advanced highly secret film making techniques like close-ups, moving camera, tracking shots and more. Up until this rewrite, I’d been framing mostly like a newspaper comic strip is framed with “square” wide-shots to show everything that’s going on at once. It took me that long to figure out what a complete waste of the medium that would be.

Using camera angles properly also opens up opportunities for humour – Pointy makes a subjective assertion supported by a tight camera angle, the next shot is a wide camera angle which makes it clear that the objective truth doesn’t match up with Pointy’s assertion at all. And hence a more engaging film.

Speaking of characters, I was putting Gronky in front of the camera and not having him do anything for an eternity while the music played. The music had to maintain interest pretty much by itself. I mean, I like these characters and they’re fun for me to look at even standing around, but they’re more fun to watch as they’re expressing their character appeal more actively – and working towards that, suddenly the structure of the film and the characters are gelling much better.

For example: Gronky used to start the short entranced by a firefly. Now we join him as he’s having a stroll. Pointy used to accost Gronky from off-screen as he was standing around, but now (after failing to attract Gronky’s attention) Pointy steps out in front of him to stop him. Between them, they’re creating situations through expressing their character instead of being placed there more artificially. Much better.

And I wrote the stupid firefly out. One less unnecessary asset to make and animate. Woohoo!

Even the plot beats have grown little notes on where the characters are at mentally – this is something I can use as notes when it comes time to get the characters moving. Just like a real film director. 🙂

Yet it’s not all plain sailing. (It never is.) I had a small hiccup today with the more organisational planning side of things. For some reason in all the books I bought and in all the websites I googled there wasn’t a succinct and well-explained guide on how best to go from a script to a master list of shots that could be used for organising storyboards. That part of the workflow, vitally important as it is for keeping things organised, doesn’t get much attention. The live-action techniques weren’t much help because live action movies are usually shot out of sequence, and for some reason the animated stuff tended to go down as far as scenes but not individual shots.

The workflow I’ve arrived at myself to deal with this problem is full of clunk and not very nice, so I’ll keep it to myself for now. Suffice it to say I have sequences, “setups” and “shots”. Multiple “shots” may share a “setup”. The images attached are the entirety of shot 9 from sequence 2 – images 9a, 9b and 9c from setup 9. That’s neat enough. The system breaks when it comes to inserting shots within stuff that’s already been boarded. “Setups” can’t be relied on for ordering because they get reused, but if I want to insert something between “shot” 9 and 10 and still have it alphabetise nicely.. or between 9a and 9b.. without renaming any files…

Anyway. I couldn’t find any information even in my collection of highly reputable books about a robust scheme for organising story beats, sequences, scenes and shots which lets a director add, alter and axe stuff without making a bloody horrible mess of the numbering scheme. I seem to remember that Titus Fehr solved the problem the way we solved that problem back in the days of C64 BASIC – leave enough room between the points of designation to allow for insertion (i.e. put your program on lines 10, 20, 30 etc so that if you need to insert something else between 10 and 20 you can put it on line 15).

Quite possibly I’m trying to do the shot planning with too much precision too early in the process as well. I’m new at this but by taking it seriously and pushing ahead with it I’m learning useful things about how to direct and produce solo. It’ll all be helpful for next time.

But wait, there’s more!

I was going to keep this a secret but since I have a better production secret to keep now I’ll come out and say it: in keeping with the “everything is unashamedly electronic” aesthetic of this project, I shall now reveal that the closing credits will be a wee tribute to the sights and sounds of the good old Commodore 64. There will be colourful paleo-digital graphics for the eyes and a rocking SID tune for the ears output from an actual 6581 SID chip courtesy of HardSID and Covert BitOps’ spiffy (and open source!) music editor GoatTracker.

I’ve already started planning out the closing music using u-he’s Zebra softsynth to imitate the instruments I’ll need before I plunge into the fun and games of GoatTracker’s wavetable-based instrument system. Based on what I’ve got so far, I’m pretty sure I can fit the tune into the traditional 3 channels using some tried and tested SID tricks but if need be I can also spread it across up to six channels.

It’s tricky to test everything I want though because many of the characteristic techniques of writing music for SID – wave sequencing, adjacent-channel hard sync and ringmod – are either impossible or plain impractical to accomplish even using state of the art softsynths like Zebra.

The SID chip, unlike all of its early 1980s contemporaries, was a hybrid analogue-digital synthesiser whose native environment was inside of a budget home computer hooked up to a programmable microcontroller with 38k of available RAM. No other synthesiser at the time let you boot into a programming environment to drive it and no other home computer of that generation had a sound chip as advanced and thoughtfully designed. What ultimately makes “sidtunes” sound the way they do is the super-fine level of control wielded by the Commodore 64’s talented musicians – control so fine that they were able to exploit a manufacturing flaw around changing volume to do sample playback.

Fortunately the control required to do a bona fide SID tune is something that trackers excel at – they aren’t as visually friendly as piano roll but when you need to work with precise numeric values instead of visual envelopes, trackers leave piano roll-based editors in the dust for sheer creative speed. I’m using GoatTracker for the final version instead of the HardSID controller running inside Reaper because it means I can control the SID from a single bespoke environment, something I’ll need if I want to use the chip to its fullest extent for making wicked cool sounds like this.

If you want to hear the mock-up, it’s up on Soundcloud.

That’s it for this week. By next Sunday evening I hope to have finished up the storyboarding and story planning.

This week 15 – 21 September I’ve been working on the soundtrack for the first time. It’s been an interesting process.

I put a rough version of the soundtrack to the storyboard and they began to differ on certain points like timing – sometimes the storyboards were right, sometimes the soundtrack was right. Even the story beats themselves started rapidly changing. It was like I’d poured three separate ingredients into a bowl with one another for the first time and they’d started an energetic chemical reaction amongst themselves. All I could do was keep making tweaks, keep rewriting music, paddle out into the surf and hold on tight.

Now that I’m doing a story reel with music – even rough music – it’s coalesced into something much more exciting and concrete. But of course the original anchor idea – Pointy stubbornly won’t leave Gronky’s shadow – is still the driving joke. Following the classic cartoon short plot formula of “set the stage then unleash the gags”, it’s still Pointy staying in Gronky’s shadow which drives the funny.

So. The soundtrack. If I ever do another music-driven cartoon – and I hope to – I’ll make a point of boarding and soundtracking simultaneously once the story beats are done. Even if it’s just an arrangement with rough instrumentation, it’ll be something to break up the monotony of other production tasks.

Composing for a music-driven animation is unlike pretty much everything I’ve ever done before – but I’m adapting. Last Sunday I spent the whole day doing an analysis of 1931’s “Smile Darn Ya Smile!” to see how it used tempo changes, motifs and rhythm; this week I’ve been looking at “You Don’t Know What You’re Doin’!” and I’ve posted my musical analysis of it online.

I’m writing a very bare arrangement right now with instrumentation that has the appropriate timbre and rhythm. Still very quick to change stuff around if some part throws off the timing – something might come in too early, stick around too long, overlap some part that it shouldn’t, etc.

My current composing-and-storyreeling workflow is something like this: I write something in Reaper (my multitrack sound editor of choice), export the result out so that I can re-time to the music in Blender; or I find something that’s suggested in the music but it isn’t timed that way in the boards, so I move stuff around in Blender instead. If the boards are really off (the images themselves are a few months old), I’ve been doing a temporary overlay in Krita. For instance, there’s a certain sequence where Gronky’s facial expression doesn’t really logically follow from the previous shot – this is because I’ve brought it ahead in the sequence of events for convenience.

So lots of bouncing back and forth. I’ve got no idea how it was done in this old days but writing the visuals and music in the rough like this is working nicely.

Moviewise I watched “Superman” (the 1978 movie) and David Cronenberg’s “A Dangerous Method”.

I don’t like modern superhero movies very much but 1978’s Superman I like a lot. There’s a confident gentleness and playfulness in it – it’s not a brain-frying session of sensory overload full of explosions, desperately showy visual effects and self-seriousness. Superman flies around being adorably awesome and sweet and aw shucks and Clark Kent is the slyest fake man-dork ever. The movie is unhurried – it takes time to settle in with the characters, let us watch the romance between Lois and Superman blossom, lets the audience soak in the sense of wonder.. and yet the spectacle when it hits in the final part of the movie is still spectacular. Superman finds out that even though he’s practically invulnerable, he can still be hurt – and emotional hurt packs the bigger sting. The movie is a lovely time capsule from before the dark and gritty superhero deconstructions of the 1980s and 1990s – and I kind of prefer it. Deathly self-serious it is not.

There are no Moment in the Sun videos this week. At all. Not a one. It’s all been soundtrack – I do have a storyboard with music but it’s not ready to show off to the general public yet. Some sort of completeness of vision will help. I did another analysis of a cartoon on Sunday night and that’s it.

In the coming week I’ll be listening to a lot of 1930s swing to figure out the arrangement, watching more old cartoons intently, writing gags for Gronky and Pointy and getting along to the next stage of production where I’m ready to start blocking out animation within the 3D program.

But first I’m going to play a ton of Wasteland 2. Man that’s a good game.