I was listening through the tracks and hit "The Lonely Pelican Has a Friend" and it absolutely wowed me with its gorgeousness combined with just a bit of glitchyness. The delay on the guitar made me think about setting up a Frippertronics-like loop and playing that over top.I loaded the original track into AUM's file player on the iPad, fired up the iOdyssey app, and found a suitable fluty/chirpy patch; I've been itching to try it out more, and this seemed like just the right thing for it.I added the K7D audio effects unit to give me a panning looper for the Odyssey, and recorded two takes of the original plus the Odyssey. Neither of which was quite what I wanted -- just a little thin...so I decided, why not try both? I loaded both takes and the original track into AUM, and played it all back in sync to record the final track. This slightly emphasized the grittiness in spots, but I decided that I liked the effect...partly because it is interesting, and partly because getting two more isolated takes was more time than I had to spend.This was definitely couch (actually bed) synthesis; I did the whole performance, recording, and mixing on the iPad. I only needed to move the final track to the Mac to MP3 encode and add the track metadata before uploading.Downloaded the final track, as noted, and tagged the file with Fission, generating a quick bit of art with ML.
I decided to pull out one of my old favorite techniques for this one: using Paulstretch with unconscionably long windows (5 seconds, 10 seconds, 40 seconds...). This takes a sample and moves it from a musical gesture to a long harmonic drone.
I started out with the ukelele track, and stretched it 2.5x with a 5 second window. I stretched the accordina 2.1x with a 10 second window to get the primary melodic and chordal motion. Both developed a melody within the drone as the stretch was done.
I then stretched the piano 1.1x but with a 30 second window, then copied it, reversed it, and duplicated this again to bring it up to approximately the length of the other two tracks.
At this point, I started layering things up in Audacity and adjusting relative levels. I decided I wanted to leave a little of the original samples audible as they were originally, so I brought in the electric guitar chords and the rhythmic bass. I realigned the chords to appropriate places against the piano/uke/accordina lines, then carefully edited the bass to duplicate one bar of the rhythm so I could repeat it to almost the end of the track.
I then took the hang sample and pitched it up 3 octaves, resulting in a funky phasing tone in the mid-register. I applied a little Valhalla Freq Echo to that to give it a little more motion, and last, I brought in the low electric bass, carefully lining it up against the hang, and duplicating notes as required to get it to the correct length.
Final mixdown was adjusting relative levels of the tracks and tweaking EQ lightly.
This was done 100% in Audacity, using the Paulstretch, fade in, fade out, and pitch shift tools, with Valhalla Freq Echo for a tiny bit of sweetening.
This started as a Ripplemaker Krell patch; I added a lot of feedback and intermodulation to achieve the trill texture, then tweaked the wavefolding and FM slightly to get some variation in the trill.
Recorded in AUM, uploaded via AudioShare, downloaded, and trimmed to 20 seconds in Audacity before re-uploading (I missed the 20-second requirement on the first go-round).
This track is a combination of old and new tools that I have been using and refining lately. Over the last few months, I've been moving more and more to modular and semi-modular synths. I've become very enamored of the West-Coast, Buchla-influenced sound and techniques, in particular the Krell patch.
I've always found "self-playing" setups fascinating and attractive, and I happened upon a Krell-like patch for MIRack that I have been extending and tweaking over the past few months, saving various versions to overlay and mix in AUM on my iPad. I also picked up Ripplemaker, which is essentially the Make Noise 0-Coast implemented in software. Krell patches on this device are even more interesting, as they can be tweaked in real-time to evolve and shift in timbre and tempo.
Today's track is a combination of five different Krell patches and Ripplemaker running yet another Krell patch and being slowly altered as the recording proceeds. This was recorded direct from AUM into Audacity (the old part) on my Mac, using another tool, the Lightning-connector audio interface to OS X.
Ripplemaker uses the Kosmonaut effects processor (another new tool!) for outboard ambience and delay effects.
The title is a riff on the "Krell Proton Field Polisher" which came up in a search for "tools of the Krell".
The hidden track is a canon by Orlando di Lasso (1532-1594); since it itself is essentially a layered loop, I decided that I'd play it out once in Garageband, build out the canon in four voices, and then find loops in the loop library that worked with the harmonies in the canon. That turned out to be a jazzy bass, two trip-hop pads, and a bossa-nova piano. I built out the "cover" loops first, then brought in the canon, dropping out cover toward the end to get one full four-layer version of the canon as the voices finished up.
I ran across this canon in 1973, in Sir Donald Tovey's "Forms of Music"; this is the first time I've gotten to actually play it!
I decided that I really wanted to play around with rhythms in the speech for this Junto. I listened to the track by itself first, then in the context of the story, and decided I still wanted to stick with that first idea.
I pulled three copies of the whole track into Live, and then processed one copy with multiple beat repeat processors (Airpusher and Semisubtle), followed by Puremagnetic's Driftmaker plugin, with Parse and Drift cranked way up). The second I used only Driftmaker on, with Chop and Drift pushed to their max).
I then experimented with various offset triggerings of the track and at last hit on restarting the beat repeat track partway through its playback as what I wanted. Because I am still a Live noob, I ended up routing it to Audio Hijack to record the different takes.
Driftmaker is unpredictable as to what it will choose to do and when, so I "played" the clips multiple times in Live, capturing the results in Audio Hijack, until I had a final take that obsessed about analysts and buttons and relatives, and captured the surreal feeling that the story had for me.
I started out trying to use the NI Bento Box and found it opaque. I couldn't get gates, routing was not transparent, and I just wasn't happy working with it.
Marc suggested setting up a VCVRack patch for people who didn't want to install the Native Instruments software. I picked out the modules that I could find that were as close as possible to the ones in the Bento Box; some were not quite as functional (the Bogaudio S&H needed an external noise source, as opposed to the built-in one in the Bento Box S&H, and the Bento Box sequencer was flexible in different ways from most of the VCV ones). After some trial and error, and a lot of staring at alternate sequencers, I had a basic set of modules.
Just to prove to myself that I should have been able to easily get what I wanted in the first place, I started patching together a basic VCO/VCA/ADSR patch, controlled by the Fundamental MIDI module. (Short pause to find and install a virtual MIDI keyboard.)
That worked fine. Well, it would be a shame not to try working the rest of these modules into the patch...and a bit later, I had a nice feedback network running; the sequencers were modifying parameters on the VCO, LFO, and filter, and different phases of the LFO were being fed back into the sequencer inputs.
The result was a very West-Coast-sounding, not-quite-repeating patch. I let it run a while, and then realized I could use the GATE from the MIDI module to reset all the sequencers. I patched that in while it was running, and then used keyboard interrupts to break up the flow and produce the coda.
“The Assignment: Make a piece of music that sounds as unlike you as you can accomplish.”
Not ambient/beatless/experimental - check No careful editing and processing of the track - check Produced as fast as possible with no edits - check Not using a traditional instrument - check Not using anything I’m comfortable with - check No programming or planning at all - check Using Launchpad, completely not my kind of app at all - check Spent less than 15 minutes total instead of agonizing over it for hours - check One take no edits - check
Pretty happy with it nonetheless! This is the Retro Funk soundpack at its default tempo. I loaded the soundpack, played with it for a couple minutes, and then hit record. Quite exhilarating having as little control as I did and cutting time-to-decision down to around a couple beats or less. Total time to make the track about ten minutes, including auditioning soundpacks. I spent more time finding and creating the art than I did playing the track.
For the art, the font is DiscoEveryday and the background vector art was created by pinnacleanimates at www.freepik.com.
Disquiet Junto 305: Three Princes
I started listening to all of the tracks; when I heard the intro to "Surathal Ranwan Samanalayo" I stopped. Birdsong, cool, got it, two more to go...and then I hit Suralo Sapa De with the beautiful call-and-response at the beginning - and I knew I was done. And going to break rule 2!
I extracted and looped the birdsong track because I had to have it as the bed to lay the other loops over, then I pulled the three call-and-response phrases and created three out-of-phase loops with those, using Audacity's tempo altering plugin to make them all exactly the same length - the original repeats only version didn't phase enough for me.
After listening to the track for a while, I decided to give it a bass line and a sustained pad to accompany the phasing loops. The pad ended up being a DX7 "sitar" pad with a lot of processing to spread it out in time and in the stereo image.
The title was selected by Google auto-complete with some twiddling of phrases, and means "Ceylon blue stone"; Sri Lanka is famous for its sapphires.
I decided to go back to an old tool for this Junto: Scape. This is a graphical composing program which runs on the iPad, and which lets you assemble generative scenes -- "scapes" -- from a large set of graphics with attached sounds and behavior. You can then create playlists that move from scene to scene, with control over how long a given scene plays and how long it takes for old objects to be removed and new ones added, but the objects themselves do what they do.
Because the individual parts interact with one another, and the parts behave differently at different times, the results are predictable only within certain limits, and the process of "composing" with Scape is primarily intuitive.
I chose to start with some relatively static textures, build them up to something much thicker and noisier, and then taper off into a more melodic texture. The initial version was okay, but not perfect; I then decided to split the track at the point I switched over to the melodic texture and combine it with the first, noisy part. The result was very satisfying and is the final track.
Translated: "There's a cat on the gamelan."
In a complete departure for me, I decided that this challenge - many different random number checks, careful timings, and so on - were much better suited to automation than trying to run them myself, so instead of an improvisational approach, I decided that I'd program this instead.
I decided that I'd use ChucK for this - it's well-suited for dynamically adding and removing layers and programatically arranging and synthesizing music.
I started out by stealing bits from the modal-o-matic example; this is a program that randomly plays different configurations of the ModalBar unit generator. The only really salvageable part (for my piece, anyway) was the code that did the different strikes and positions on the virtual bar. Playing around with this, I decided to use a single note (middle C) with different strikes, positions, and presets on every note as an ostinato, and then create dynamic "shreds" that would compute the note timings and durations according to the original plan.
After a few hours of programming, mostly figuring out the timing, I was about to record the final version, when I realized that because this was ChucK, I could actually run six independent copies of the piece at the same time; a couple of takes (every take is different!) and I had the final six-layer version.
Edit: I realized on listening once I was awake that the multiple layers of pulse were obscuring the panning; split the pulsed and no-pulse versions apart and ran six of the no-pulse and one of the pulsed for the final version.
Edit edit: Soundcloud's conversion is sadly covering up the panning in the pulse track. I suggest running it yourself with headphones to get the intended effect.
https://github.com/joemcmahon/ada-kucing-gamelan.ck
More on this 235th weekly Disquiet Junto project — “Create a piece of music based on a structure determined by the roll of a single die” — at:
http://disquiet.com/0235/
More on the Disquiet Junto at:
http://disquiet.com/junto/
Join the Disquiet Junto at:
http://soundcloud.com/groups/disquiet-junto/
Subscribe to project announcements here:
http://tinyletter.com/disquiet-junto/
Disquiet Junto general discussion takes place on a Slack (send your email address to twitter.com/disquiet for inclusion) and at this URL:
http://disquiet.com/forums/
I listened to the first thirty seconds of each of last week's submissions, and ended up picking "Tired Machines", "Hexhead", and "Il trio mutilato" as my sources, and I decided that I'd use Hexhead's signature motif as the focal point for my track.
I clipped off the first thirty seconds of each source track in Fission, and further snipped up the Hexhead and Tired Machines clips into two and three smaller clips, respectively, to capture the specific musical gestures in each.
I then tossed these into the shredder with Forester, doing several takes until I got some source materials I really liked. I chose to use the "clumped" trees, and mostly used manual placement of the playhead; there were a couple of short segments where I allowed the head to wander at random. The shortened clips made it easier to get output that featured twisted echoes of each of the source tracks.
I loaded both the resulting takes into Audacity and abused noise reduction to thin out and distort the first take by selecting a dense but not very interesting section of the track and using that as the noise to be removed. I then applied that noise profile to the other take to get some crunchy spectral distortion.
I then selected sections of both takes and used repeated applications of Audacity's sliding tempo/pitch plugin to stretch and compress them, and mixed the two takes together into a final track, adding some silence to line everything up.
More on this 234th weekly Disquiet Junto project — “Make a remix of three tracks of a remix of three tracks, courtesy of a Creative Commons license” — at:
http://disquiet.com/0234/
Be sure to credit all the tracks (first and second generation) employed in your piece.
More on the Disquiet Junto at:
http://disquiet.com/junto/
Join the Disquiet Junto at:
http://soundcloud.com/groups/disquiet-junto/
Subscribe to project announcements here:
http://tinyletter.com/disquiet-junto/
Disquiet Junto general discussion takes place on a Slack (send your email address to twitter.com/disquiet for inclusion) and at this URL:
http://disquiet.com/forums/
Thanks to Audio Obscura (aka Neil Stringfellow) for proposing this week’s project.