Although last year I immediately made a patch once Jamuary was over, this year was a little different. Jamuary really wore me down, and I just wasn’t ready to step back up to the synth. A patch a day everyday for 31 days is a lot of patching. They all seem to run into one another in a huge blur. But as tiring as Jamuary was this year, I knew it wouldn’t be very long before I had an itch to scratch, and this one came suddenly.
It’s amazing sometimes where one might get inspiration. Sometimes it’s a bird singing or the sight of something special. This time, for me, was the last place I might have thought. A short snippet from Sarah Belle Reid’s video documenting a Buchla 200 system, one of the first to have been built, wherein she did a patch using two smooth random voltages to modulate the wave shape of the Buchla 258, and two unsynced cycling triangle functions to control their respective levels. There was nothing fancy about the patch, nor was it particularly musical in context. It wasn’t designed to be anything more than a very simple demonstration of the 258’s waveshaping abilities. But something about that sound really captured my attention. I listened to that section several times, and decided that was my next direction.
Sketch 21 started with two Contour envelope generators cycling unsynced. I’m not sure why I went with Contour, a discontinued take on an ADSR envelope generator, when I have several AD function generators in the Make Noise synth that are quite like the Buchla 280 Quad Envelope Generator used by Sarah. Perhaps it was because I had just received my second Contour to complete a pair, as well as the “legacy” section of my Make Noise Synth.1 My plan was to use two Make Noise Classic workhorses DPO and STO, after all. Or perhaps I thought I might use the Sustain parameter on Contour. In any case, I decided to use it and quickly discovered one of its “shortcomings.” Contour is old school. Yes, it can cycle, but you have to do it via patching the End Of Cycle output to the Gate input. There is no “Cycle” button or switch like there is on Function, Maths (at least the modern Maths), or Polimaths, which necessitated manually triggering each Contour using a PrssPnt and quickly switching patch cables in order to get the patch started. This wasn’t a problem, per se, and it’s not really a “shortcoming” that Contour needs manual triggering to start a cycle. The Buchla 280 functions in exactly the same way, but it is a conscious step that must be affirmatively taken, and it needs to be executed properly in the right time frame lest you be forced to start the patch over again. Fortunately both envelopes were quite long, approximately 10 and seven seconds long at the start of the patch, but I won’t say I was perfect on every run through. Once I set this part of the patch, however, I was pushed into changing it slightly. Contour is a Make Noise version of an ADSR. In order to sustain notes, a gate is required. But the End Of Cycle output on Contour only outputs a trigger when it’s patched to cycle. To get around this small hurdle, I patched the EOC gate outputs on the Contours to René’s X and Y clock inputs, with both channels set to output gates on each step, which then gated each envelope as it cycled. As each envelope completed, the EOC trigger moved René, which in-turn gated Contour to restart the envelope, with the benefit of a gate to use for the Sustain parameter. That said, I’m not really convinced that those gates were long enough to activate the Hold parameter. I couldn’t really hear a palpable difference, but stuck with the patch.
The envelopes created were fairly long, about 10 seconds and 7 seconds total, with an Onset (read: the A and D in an ADSR) long enough to hear the volume move in, as opposed to having a hard transient, and a long Decay. The timing of the envelopes moved the sequencer, but the functions themselves also controlled the articulation of each of the two oscillators used in the patch. I regularly use slow moving waves to control oscillators like this. Like, a lot. Most often they have rather long attacks, but I wanted the attack of these envelopes to be more like bow strokes than long fades. Smooth but not overtly long. A noticeable rise, while still exhibiting some measure of haste.
Once the envelopes and their general cadence were set it was time to move on to the two voices. My first thought was to really modify this patch by using Spectraphon and thinking about waveshaping via modulating its Partials, Focus, and Slide parameters, but I quickly moved away from it. Instead I took the route, in Sarah’s footsteps, of using oscillators with waveshaping as an explicit parameter, choosing a mix of the sine and Final outputs on DPO for the bass, and a mix of the sine and Shape output of STO for the higher voice. In Sarah’s video, she very specifically used a Smooth random output to modulate the wave shape for each oscillator of the Buchla 258 so that the timbre and tone was constantly shifting in unpredictable ways. That was the entire point. And while I initially followed Sarah’s patch by using the Smooth output from Wogglebug, by the second or third recording I had switched to using a modulated chaos signal from Spectraphon, via a Spread Multimod. Throughout most of the six recordings of this patch I used four of these Multimod outputs to modulate the waveshaping of both oscillators. Shape, Angle, and Fold on the DPO, and Shape on the STO. But something felt odd in the modulation on the DPO. It would too often cause the bass to all but disappear because of how Shape and Angle interacted with one another. This is, of course, expected, but it had worn out its welcome after several iterations and constant adjustments to try and mitigate the issue. Once I arrived at this final revision I was using DPO’s Oscillator A to modulate the B side, via the Mod Bus, at audio rate. The results are a richer bass throughout, with sometimes wild timbres and textures. I stuck with an attenuated, moderately slow moving chaos signal from Multimod to modulate STO’s shape, which, to my ears, creates some of the most interesting movement in the piece.
As each oscillator was introduced, both started as humble sine waves. Even as gain was added, the bass note never seemed to lose that “saturated sine” sound, but the higher voice from STO was very woodwind like. At times one might hear something like a clarinet or a bassoon. But as the piece progressed, I introduced the Final and Shape outputs via mixing in a pair of modDemixes. Each mix output was sent to QMMG’s channels one and four respectively. I tried every sort of amplitude processing I had in this synth. My initial thought was to, like Sarah, use vactrol controlled low pass gates. Then I experimented with using a low pass filter with the cutoff fully closed, but that wasn’t as nice. Then, on a whim, I tried modulating the level in regular old VCA mode. To my surprise, this allowed for the sort of response I heard in my head, and I stuck with it (though I did revisit these experiments at least once more through the tweaking process). Alongside the oscillator outputs I also introduced a form of noise2, courtesy of the Woggle output from the Audio section of Wogglebug. Each noise source was enveloped in a DXG using the same Contour envelopes controlling the rest of the audio, before being high pass filtered through QMMG’s channels two and three using inverted versions of those same envelopes. As audio from each oscillator fades in, so does the noise, and as it gets louder, it also gains a bit of low(er) frequency information to fill out the sound.
All four sounds, the oscillators and noise, exited QMMG via the mix output before being sent through Bruxa for even more heavily texturizing of the sounds. As the patch began, Bruxa’s input level was set moderately with no distortion. As the piece progressed I added distortion to both voices via QMMG’s feedback knob, as well as by eventually pinning Bruxa’s input level before introducing the wet signal. This allowed for a slow build and lots of harmonics to chew on in Bruxa’s filtering and delay processing. As the recording progressed I slowly raised Bruxa’s mix to full wet, allowing its cryptic feedback loop(s) to introduce its own chaotic behavior. But I also wanted more movement, and decided to repurpose two of the chaos signals from Multimod that had initially modulated DPO by using them to modulate Bruxa’s Decay and Absorb, while modulating Time with one of Bruxa’s CV outputs. The resulting sounds are huge, oftentimes almost cinematic in scale.
Bruxa’s output was sent to the output mixer (AUM via an ES-9), where it was stereo-fied by a brilliant plugin called Haaze. Though at its most basic level Haaze does use the Haas Effect, it has a full cabinet of tools one might use to create a stereo signal from a mono one, or liven up a pedestrian stereo signal. It’s a subtle yet impactful way to process a mono signal to create as wide a stereo field as you want. I chose to remain a bit more modest. The stereo field is there, but not overtly wide.
This signal was then sent out of the mixer to Mimeophon for some stereo delay. The third repurposed chaos output from Multimod was patched to color just to create a bit more subtle movement. Everything was sent to the Maneco Labs Otterley for reverb. And what was interesting is I finally found a way to utilize its Reverb 2 as one of my favorite effects: granular delay. These sounds leave a trail of pitch shifted grains in the wake of a sound like a bio-illuminescent trail in the water. At times they sound like an accompanying string orchestra echoing back with tremolo strokes in an intensely beautiful way.
If I’m being completely honest, this recording wasn’t originally conceived of as a normal post on peaks and nulls, but as a recording for something to be released. But I was never able to quite capture the magic of this patch in a single performance. Feel-wise, rev2 is by far my favorite of the recordings. Everything seemed to line up perfectly. The modulation was at the perfect speed. The interactions between modulated signals was ideal. But that was a very early recording before a lot of tweaks were made. Most of those changes improved the recordings immensely, but I see this patch as one that got away. Something was lost amongst those revisions, never to be recovered.
Modules Used:
Contour
René Mk2
DPO
STO
QMMG
Bruxa
Maths
Spectraphon
Multimod
modDemix
DXG
Wogglebug
Mimeophon
CV Bus Mk2
Maneco Labs Otterley
Plugins Used:
Klevgrand Haaze 2
Klevgrand Luxe
Toneboosters Equalizer 4
Performed and recorded in 1 take in AUM on iPad via the Expert Sleepers ES-9.
- Although not all of the modules in this section are officially discontinued, some of them are (Echophon, Contour), while others are all but discontinued, not having been made in several years (Optomix v2, LxD), and the DPO is an OG DPO with 6 vactrols (last made in 2014 or early 2015). The only readily available modules in the row are Maths and STO. ↩︎
- Count me among those who would do unspeakable things for a simple white noise output in Make Noise’s catalogue that doesn’t require patching and using complicated modules. It’s noise! Even Buchla and Serge thought it appropriate to provide a noise output. ↩︎
