This was an experiment in exploring the device as a whole: not just as a set of loops contained in one object, but the sequence of those loops, and the way the loops progress as a result of the object’s design. All the previous Buddha Machine Variations to date have involved one loop per machine. Here the multiple loops of such a device, one from the second-generation of Buddha Machines, are heard not just in sequence, but in overlapping sequences. They’re considered as a suite.
There are three subsets of whatever loop is playing at a given time. Each is a subloop: one three seconds, one four seconds, one five (in the ER-301 module, using the Feedback Looper unit). Each of the subloops is extracted from a different narrow band of the audio spectrum (via the FDXf module). As the piece proceeds, the button on the side of the Buddha Machine is clicked. One or more of the subloops begins recording the new audio almost immediately, but there is some time before the preceding loop is entirely eradicated. At the very end, the audio cable is pulled out of the Buddha Machine, and eventually all the subloops give way to silence. (This was less sudden when rehearsed, but it works OK here. I’ll be trying this all again with a slightly different approach.)
For further patch-documentation purposes, here is a shot of the synthesizer:
One loop from the fifth generation Buddha Machine, little snatches of it glitching out in a semblance of correlation. At first all that’s heard is the unaltered audio of the source loop. That’s the fifth channel of the mixer (lavender cable). And then one by one, the first through fourth channels of the mixer are introduced. Each is set on a mico-looping procedure of the source loop, varying individually between roughly 100 and 300 milliseconds. They are each set to record over whatever was there before, meaning sometimes there will be silences. And finally, once the main source audio is turned off in the mix (at 1:45), all that is heard is this combination of microloops.
If you’re familiar with the ER-301 module, here’s some additional detail: The unit being used is the Feedback Looper. There are four of these running individually. Each is taking a square wave from the Batumi as its on/off switch to record. In addition, each is taking the next channel’s on/off as its trigger to (un)engage (the fourth channel looks to the first channel, squaring the circle).
That runs for quite awhile, and then at 2:55 the relative pace of the four square waves are altered. This has an impact on when the individual loopers are triggered, and then at 3:09 the Batumi is switched from “free” mode to “quad,” and the levers adjusted further.
For further patch-documentation purposes, here is a shot of the synthesizer:
One loop from the first generation of the Buddha Machine, which turns 15 years old in 2020. The loop ends up with three variants in the mixer, channels one, two, and five. The fifth channel (lavender cable) is just the straight audio out of the Buddha Machine (a little noisier than usual, perhaps because the batteries were going). The first channel is the output of the Muxlicer, which switches between eight variants on the audio: six are spectral bands (from the FXDf), and two are channels of the granular synth (the position of which, in the Smog, is shifting regularly). The Muxlicer is going at a steady pace (clocked by the Dixie II, starting when the Muxlicer’s switch is flipped at :28), but the sequence is random thanks to an inbound sine wave (from the Batumi). The second channel, which kicks in at 3:24, is a tiny fragment of a loop that is constantly being overwritten, and it’s set off from the Muxlicer’s beat by a slight, almost half-second delay, which at first gives the contrast a sing-song quality, but eventually disappears as the amassed swell occludes virtually everything. And that’s about it, except that at 4:25 a slow wave is introduced to gently alter the volume of the original, channel-five track.
For further patch-documentation purposes, here two shots of the synthesizer:
This is a Saturday evening trial run at selecting tiny bits of a single loop from a Buddha Machine and having them pulse against each other. Eventually the main source line is eliminated so that only these microloops remain. Toward the end, two of the microloops are further affected by delays that warble by changing ever so slightly within a narrow range at a rapid pace. There’s not much more to it than that. The source audio is the third generation of Buddha Machine, called Chan Fang, which came out a decade ago, back in 2010. The instrument heard is the qin, a Chinese zither.
There are five channels into the mixer. The purple one, channel five, is the unadulterated source audio. Channels one through four each get their own individual micoloop in real time, one at a time. Channels three and four are the ones eventually set to warble (channel four starting at 3:24, channel three at 4:52). The loops and delays occur in the ER-301 (the large synthesizer module in the lower left, running the Pedal Looper and Delay units). The warble is a result of a pair of waves that come from a combination of the Batumi (the one with four vertical levers directly above the ER-301) and the S.P.O. (the one with four black knobs, two to the right of the Batumi). I’m going to explore this approach further, likely combining microloops from various loops of the Chan Fang.
For further patch-documentation purposes, here is a straight-on shot of the synthesizer:
One thing that occurred to me as I’ve reached the end of a full month of daily Buddha Machine Variations is that I am accessing a familiarity with the source recordings that has only one direct comparison for me, which is the year I spent daily listening to and writing about one of my favorite albums of all time, Selected Ambient Works Volume II by Aphex Twin. As I searched for a Buddha Machine loop for today’s exercise, I realized that as I clicked through the options, I knew what would be next: Not only were the individual loops familiar as a friend’s voice on the phone, so too were the loops’ relative proximity to each other. And as I listened to the audio as it was being processed, I adjusted the various options in accordance with the specific loop’s qualities. And when I thought back on how this patch came together, I realized the decisions about what kind of processing to engage in were rooted in the source material. That point is distinct from where I began 31 days ago, when the audio was more of a discreet object, and the processing a discreet application. The divide between them has narrowed.
In any case, this piece has four subsidiary parts. I’ll describe them in brief, one channel of the mixer at a time, from left to right. Channel one is the straight audio out of the Buddha Machine, but with a slight variation on the delay, lending it a warble quality (delay via the ER-301). Channel two is the same unadulterated audio, its volume increasing and decreasing thanks to a slow, shallow wave (courtesy of the o_C module, running the Hemisphere firmware). Channel three is a narrow band of the FXDf-derived audio (from the source loop), put through a slight, static delay (in the ER-301). And channel four is another narrow band of the FXDf-derived audio, set on a little loop that is constantly overwriting itself. The loop is a little over two seconds long, and the trigger for the recording is going at a different pace, so over time the audio that is heard in this loop changes. (The loop kicks in at 1:05, when I connect a cable from the Batumi to the ER-301.) Beyond that, it’s all just me manipulating the various channels, based on my understanding and appreciation of the source loop, its melodic qualities, and the stages that comprise its structure.
There was no entry yesterday because after I recorded it last night I realized there was an issue with the audio, so I waited until today to re-record it. In the process I changed the patch, adding some elements.
For further patch-documentation purposes, here are two shots of the synthesizer:
And the settings on the Ornament and Crime module, the screen of which had gone dark while the video was being shot:
Marc Weidenbaum founded the website Disquiet.com in 1996 at the intersection of sound, art, and technology, and since 2012 has moderated the Disquiet Junto, an active online community of weekly music/sonic projects. He has written for Nature, Boing Boing, The Wire, Pitchfork, and NewMusicBox, among other periodicals. He is the author of the 33 1⁄3 book on Aphex Twin’s classic album Selected Ambient Works Volume II. Read more about his sonic consultancy, teaching, sound art, and work in film, comics, and other media
• July 28, 2021: This day marks the start of the 500th consecutive weekly project in the Disquiet Junto music community.
• December 13, 2021: This day marks the 25th anniversary of the founding of Disquiet.com.
• January 6, 2021: This day marks the 10th anniversary of the start of the Disquiet Junto music community.
• There are entries on the Disquiet Junto in the forthcoming book The Music Production Cookbook: Ready-made Recipes for the Classroom (Oxford University Press), edited by Adam Patrick Bell. Ethan Hein wrote one, and I did, too.
• A chapter on the Disquiet Junto ("The Disquiet Junto as an Online Community of Practice," by Ethan Hein) appears in the book The Oxford Handbook of Social Media and Music Learning (Oxford University Press), edited by Stephanie Horsley, Janice Waldron, and Kari Veblen. (Details at oup.com.)
• The Disquiet Junto series of weekly communal music projects explore constraints as a springboard for creativity and productivity. There is a new project each Thursday afternoon (California time), and it is due the following Monday at 11:59pm: disquiet.com/junto.
Since January 2012, the Disquiet Junto has been an ongoing weekly collaborative music-making community that employs creative constraints as a springboard for creativity. Subscribe to the announcement list (each Thursday), listen to tracks by participants from around the world, read the FAQ, and join in.