Category: studio journal

So, to cut to the chase, there’s a big whopping error in the middle of this video. Actually, earlier than the middle. At 1:32 (the full length is 4:37), I connect a cable, and then proceed, over the following minute and a half, to 3:10, to first recognize something is off, and to then sort it out. I had the whole thing plotted, but it had been a long day and I was pretty tired, and I messed it up when I started recording. Still, that minute and a half sounds fine, and even as I was sleuthing what was wrong, I was thinking, “This actually sounds fine, so long as I can fix it soon.” As further evidence I was a bit rushed, the Buddha Machine that is the source of the sound is entirely off-screen. It’s the Chan Fang model, which is all samples of a performance on a qin, an ancient Chinese zither.

Three days/patches ago I mentioned parenthetically, of the big module in the lower left: “Of course, there’s a lot going on inside the ER-301 that’s not viewable.” If you watch that module from 1:32 to 3:10, you’ll get a sense of what moving around its interface is like.

Let’s break the recording down. There are four channels in the mixer, from left to right: one/lavender, two/black, three/yellow, and four/white. They are all variations on the incoming Buddha Machine loop.

Lavender is just the sound of the Buddha Machine loop, unaltered (putting aside the extent to which slight changes in the gain on the initial input do have an influence on the sound). Black and yellow are processed by the ER-301. Both are variations on bands of the incoming audio split off from the overall audio spectrum. Black is on a short delay, so it echoes a bit of the main signal (listen at 1:04 to how it repeats what just preceded it, as if a full register lower).

Yellow is the problematic channel. It is on a two-second loop that is constantly writing over itself when triggered. The signal that triggers the loop is itself fairly long, so the yellow signal doesn’t change that much. The problem was, it’s set in the ER-301 to be triggered by the gate on the top row, and I had it set on the second row. Eventually I figure this out, which is why at 3:10 I change the location of that cable, and at 3:16 the looping kicks in.

As for white, the fourth and final channel in the mixer, it is another band of the audio spectrum, sent through a granular synthesizer. And that covers it.

For further patch-documentation purposes, here’s a straight-on shot of the synthesizer:

Video originally posted at youtube.com/disquiet. There’s also a (https://www.youtube.com/playlist?list=PLAgCxRbmR1MIM4mCYe17nERi9xeEWAD2w) of the Buddha Machine Variations.

One loop, three strands, many rhythms. A single loop comes out of the second-generation Buddha Machine. It’s split into two. One cable goes straight into the mixer. That’s what’s heard when the first pot is turned up. The second cable goes into a filter bank, two cables from which go into the ER-301. Each carries a different band of the audio spectrum of the source loop. Each of those goes through a different delay in the ER-301, and then the volume of each of those fluctuates (using the ER-301’s Linear Unipolar VCA unit) due to waves coming from the Batumi. The ADDAC mixer has three options for each channel: solo, off, and mute. At the end, each of the three mixer channels is turned off one at a time.

For further patch-documentation purposes, here’s a straight-on shot of the synthesizer:

Video originally posted at youtube.com/disquiet. There’s also a (https://www.youtube.com/playlist?list=PLAgCxRbmR1MIM4mCYe17nERi9xeEWAD2w) of the Buddha Machine Variations.

This reworking of a Buddha Machine loop is based on small, slowly accruing sub-loops (subsets of the source audio), much as was yesterday’s. There are some differences. The samples going into the buffers are more likely to cancel out the audio that preceded them, rather than build to an orchestral density (the change suits the source loop, from the fifth-generation Buddha Machine, which has a genteel, chamber-music quality). The sub-loops also each fade in and out, rather than cutting at the seam of each new sample. (Also, yesterday’s experiment used a different generation of Buddha Machine.)

When the video starts, all that’s heard is the loop itself, a combination of stately, simple piano and strings. At 26 seconds or so, a new element enters, not fully heard until about 38 seconds: it’s the source audio sliced and reworked courtesy of the Muxlicer. What’s going on is that as the audio comes through, there are eight potential ways we’ll hear it. Half of those are the audio, with the relative volume adjusted a bit (note the red faders above each of the eight jacks). Two are bands of audio extracted from the spectrum, courtesy of the FXDf module. And two are the left and right channels of the source audio going through a granular synthesizer. The pace of that Muxlicer is set by the pulse going through the lavender patch cable. The resulting Muxlicer activity goes into the first channel of the mixer (white cable).

At 1:13, a new element enters, the buffer looping. This is all happening in the ER-301 module (the Feedback Looper unit). What happens is that the source audio is being fed into a five-second loop, which itself is overwritten with small bits based on the pace set by a square wave from the Batumi (this goes into mixer channel two, yellow cable) . At 1:36, a second buffer loop enters, this one on a six-second loop (mixer channel three, blue cable). Eventually, around 3:18, the Muxlicer is turned down fully, so we only hear the contrasting buffered loops, until it all fades out.

For further patch-documentation purposes, here’s a straight-on shot of the synthesizer:

Video originally posted at youtube.com/disquiet. There’s also a (https://www.youtube.com/playlist?list=PLAgCxRbmR1MIM4mCYe17nERi9xeEWAD2w) of the Buddha Machine Variations.

A lot of these Buddha Machine Variations have been static/semi-generative things that just proceed under their own power. This is, I think, only the second where I intrude and do something to initiate changes.

The source audio is a loop of a Philip Glass choral piece emanating from the edition of the Buddha Machine created by FM3 to commemorate the minimalist composer’s 80th birthday.

The first time through, the loop is heard on its own. The second time through, beginning at about 30 seconds in, what’s heard is two treatments of the loop together, along with the original loop. You see me, during the pause, turn up two of the pots on the mixer, and turn down slightly the one that had been playing. The simpler of these two treatments is simply a narrow band of the audio, extracted from the spectrum.

The other treatment is a little more complex. It’s the Muxlicer (black module with eight vertical sliders in the upper right) clicking through at a steady pace. The module has eight output stages. There’s a main line, and then you can slot in alternates for any of the eight stages. When it’s first heard here, two are playing additional bands of the audio spectrum, which is why those moments stand out.

If you keep an eye on the horizontal lights across the Muxlicer, you can see a little red one pop on to show which channel is playing. The pace at which the lights change is steady. What is random is the sequence, as regulated by a sine wave (from the Dixie II module). It was interesting trying to find a pace that contributes the desired amount of randomness, when I was preparing the patch.

That’s the second time through. The third time through, again during the pause (at 60 seconds), I turn down the original line entirely, leaving just the two treatments that appeared in the previous round. The step-by-step pace is more prominent as a result.

Then I do two more things: I plug first one and then, at about 1:37, a second cable into the ER-301. That’s the large module with two screens in the lower left. What’s happening in the ER-301 is that the full audio of the original loop is, itself, being cut up and looped. There are two channels out of the ER-301 (this is using the Feedback Looper module). One is recording a four-second loop over and over, and the other a two-second loop. The recording of the loops isn’t in sync with the playing, so the snatches overlap haphazardly, piling overtones. If you were to watch the segments accrue, it’s a little bit like a fast motion audio version of posters being layered on the wall of a construction site.

For further patch-documentation purposes, here’s a straight-on shot of the synthesizer:

Video originally posted at youtube.com/disquiet. There’s also a (https://www.youtube.com/playlist?list=PLAgCxRbmR1MIM4mCYe17nERi9xeEWAD2w) of the Buddha Machine Variations.

The goal here was to push the Muxlicer by altering its clock as time passes. The clock is what sets the pace of the changes that you hear. The Muxlicer, from the manufacturer Befaco, is the module toward the upper right corner. It’s the one with all those red faders that go up and down. The way the Muxlicer works (in this scenario, for there are various others) is you put in one master audio line that is then spread across the eight stages, the relative volume of which are individually set according to those faders. And then you can slot in alternate audio in as many of those eight stages as you like. So, as the clock clicks through its paces, you hear each of the eight stages in sequence, with the alternate audio occasionally replacing the main audio, and the volume of each stage shifting as well.

I’ll get to the source audio itself in a moment, but first about that clock situation. Initially while patching this, I just had a steady clock coming from a trigger pulse from the Dixie II (manufactured by Intellijel). If you want to know what the Muxlicer’s clock pulse is, keep an eye on the rate of the little light toward the top of the Dixie II, which is the seventh module in from the left on the top row. To give the steady pace some vibrancy, I had another wave, a sine from the Batumi (from the manufacturer Xaoc), create a random-ish, ever-shifting sequence as to which of the eight Muxlicer outs would play at any given moment. So far, so good.

But I wanted to push the clock further. This effort came in two phases of the patch’s development. The next thing I did was feed a slow moving square wave from the Batumi into the pitch control of the Dixie II. This meant every time the Batumi wave was on, the pitch of the Dixie II would suddenly level up, meaning the clock signal sent to the Muxlicer would speed up. This worked well, but an interesting thing happened. Since the square wave initiated this change in regular increments, going back and forth, the change in speed of the Muxlicer fairly quickly came to feel rote. If, as Brian Eno said, repetition is a form of change, then this was an example of how change (the clock going from fast to slow to fast, round and round) became a form of repetition. Something needed to change, in other words, for this to feel less repetitive. Something else needed to make that change occur.

I should note that the square wave going into the pitch of the Dixie II is squashed in height a bit by the S.P.O., from the company Steady State Fate. Initially this decision was made in order for the change in clock pacing to not be too significant. But then something occurred to me: the S.P.O. can combine signals, so I put a second, more chaotic wave also into the S.P.O., meaning that every once in a while, the square wave affecting the speed of the clock out of the Dixie II would itself be altered with a rapid little uptick. As a result, the change in the rhythm evidenced by the Muxlicer became more complex. If you keep an eye on the Muxlicer, there are eight horizontal red lights just below the faders. Watch those, and you’ll see how the clock changes. Where that additional, chaotic wave form comes from is it’s an envelope from the Detect-Rx, which is following the volume of one of the bands coming out of the filter bank. Which is to say, there’s some chaos going into the clock, but at some level it is correlated with the sound of a Buddha Machine itself

There’s more going on regarding the audio inputs. The blue one is first generation, and the peculiar color is second generation. The blue one (with the green audio cable) is split, and half the signal goes into a filter bank while the other half goes into a granular synthesizer. Three bands come out of the filter bank. One goes into the main Muxlicer in, one into an alt-in, and one into an envelope follower (more later). The left channel of the output of the granular synthesizer also goes into the Muxlicer alts (twice). The right channel of the granular synthesizer goes into a second channel of the mixer. Its volume is then altered thanks to a square wave coming out of the Batumi. However, I wanted to find a complement to the sudden jerks (the shudder) of the clock, so I put a burst-like signal into the trigger of the granular synth. The signal is a square wave, the frequency of which is being altered by a different square wave, the frequency of which is itself being altered by a sine wave (all within the Batumi; I believe this is referred to as self-patching). If you look at the lights of the four red faders of the Batumi (fourth module in from the upper left), you’ll see when that rapid burst kicks in. And, of course, you’ll hear it.

Which leaves a question: where’s the second-generation Buddha Machine? It’s not heard here. A patch starts in one place, and ends in another. At some point I disconnected the second Buddha , was pleased with the result, and never patched it back in. I didn’t fully recognize this until I was done.

For further patch-documentation purposes, here’s a straight-on shot of the synthesizer:

Video originally posted at youtube.com/disquiet. There’s also a (https://www.youtube.com/playlist?list=PLAgCxRbmR1MIM4mCYe17nERi9xeEWAD2w) of the Buddha Machine Variations.