Probabilistic Sequencing with STG VMS + Dotcom
Aug 22, 2010 Uncategorized
Okay, fancy titles aside, this post is about allowing you to do “probabilistic sequencing” using synthesizers.com modules and the STG Soundlabs Voltage Mini-Store (two of them, actually). You can probably use this with any sequencer you may have, but I have the VMSes, so that’s what I used.
Definitions first; What I mean by “probabilistic sequencing” is this: In deterministic sequencing, you might use a set of knobs to set voltages and a set of switches to set gates for each step. A knob sets the pitch of the note to be played at that step, and a switch determines whether or not that step will be played (to allow you to have rests in your sequence). What we’re doing instead with probabilistic sequencing is to have a pair of knobs for each step. The first one determines the pitch being played at that step, and the second one determines the likelihood that that note will be played. If the knob is twisted completely counter-clockwise, you get a rest. If it’s twisted completely clockwise, you get a note. If it’s somewhere in between, you may or may not get a note, and the likelihood is in proportion to the position of the knob.
This was actually sparked by a question from Mark Sims on the synthesizers.com Yahoo group, but I became interested because I use probabilities and randomness extensively when sequencing in Nodal, and I find the results very effective, so being able to do that in hardware seemed like fun.
My solution uses the following modules: 1x Q110 Noise, 1x Q117 Sample & Hold, 1x Q125 Signal Processor, 1x Q124 Multiples, 1x Q128 Switch, 1x Q109 Envelope Generator, 1x Q108 Amplifier, 1x Q106 Oscillator, 2x STG Soundlabs Voltage Mini-Store.
The solution that I came up with (which is a slightly different arrangement than the original question, so it’s been modified for my configuration) looks like this:
I don’t know if that’s the best diagram ever. If you have any comments on the documentation style, let me know. The patch in words looks like this:
NOISE (White Noise) -> SAMPLE & HOLD (Input)
SAMPLE & HOLD (Output) -> SIGNAL PROCESSOR (Top Input)
SAMPLE & HOLD (Gate) -> MULT B
SIGNAL PROCESSOR (Top Output) -> MULT A
MULT A -> SWITCH (Control)
MULT B -> SWITCH (2A)
MULT B -> STG VMS 1 (Shift)
MULT B -> STG VMS 2 (Shift)
SWITCH (1 Common) -> OSCILLATOR (1V/Oct)
SWITCH (2 Common) -> EG (Gate)
OSC (Square Out) -> AMP (Signal Input 1)
EG (Output) -> AMP (Control Input 1)
AMP (Output) -> Mixer
STG VMS 1 (Output) -> SWITCH (1A)
STG VMS 2 (Output) -> MULT A
Obviously you can customize this to your liking. For example, you might want to choose a different waveform, or mult the pitch out to multiple oscillators or something else.
Okay, so what’s going on here?
The first Voltage Mini Store in this set up is being used traditionally, to set pitches. The second is our probability setting for each note.
Well, the Noise source is basically operating as a random number generator. The Sample & Hold is running the timing of the sequencer, both by grabbing just one value from the noise source per beat and by using its gate output to manually shift the sequencers. I left the output and input knobs on each of these at full. The frequency knob on the Sample & Hold will be set to taste for the speed of your sequence, and the timing source of the Sample & Hold is set to Internal.
The multiple group called MULT A in the description above is being used to merge our random number generator with the output of our knobs from the second Voltage Mini Store, and send that to control the Switch. The output from the Noise and S&H goes through a Signal Processor en route because you’ll have to carefully set the values of the gain and offset until you get the probability curve that you want. The method for doing this is basically to set all of your knobs to 50% and then adjust until they fire about half the time, then check to see if they fire all the time at 100% and never at 0%. If not, adjust accordingly until you get the response curve you want. It sounds fiddly, but it didn’t take me very long to get something workable for me. You’ll probably want to slow down your timing while doing this to make counting easier. (I also cheated by temporarily unplugging the shift inputs on both Voltage Mini Stores so that I could just fiddle with one knob instead of having to twist them all to try the various values.)
I send both the pitch and the gate through the Switch because if you decide that you want a log decay tail on your envelope and you don’t do this, the decay tail will change to the pitch of the next note in the sequence even if that note isn’t supposed to be sounding.
From there it’s pretty straightforward. The pitch drives your oscillator, the gate drives your EG, they go to the Amp and Bob’s your uncle.
Here’s a short recording of the patch in action. In this example I’ve got notes 1, 3, 5, 7 at 100% (they always sound) and the alternate notes 2, 4, 6, and 8 at 50%, so you’ll hear them fill in sometimes and sometimes not. I’ve also added a bandpass filter sweep and some reverb because I can’t leave well enough alone. You’ll notice that there’s a little rhythmic click after each note. That’s not supposed to be there. I kind of like the way it sounds, but if someone has suggestions for getting rid of it, I’d love to hear it.
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
Justin Bieber Redux
Aug 17, 2010 Uncategorized
So, as you all knew I would, as soon as I got home, I downloaded that slowed-down Justin Bieber track and sped it up 8x to see what we got.
They’re not lying. It’s definitely the Bieber song. However, a lot more has been done to it than just slowing it down 8x.
At first I thought maybe it was a different version — perhaps an “unplugged” version. I didn’t listen to the original and the re-sped-up version side by side enough to say, but I’d posited that it was a different rendition that felt a lot more soulful and bluesier to me. However, from reading the comment thread on the post, it appears that it’s just that it was actually slowed down somewhere in the neighbourhood of 10.7x instead of 8x, and that in addition it was knocked down in pitch by a minor second, which presumably account for the change in tone and feel to my ears.
Also, the shimmery, dreamy quality seems to come from a large hall reverb applied to the original prior to slowing it down. You can definitely hear it when you speed it back up, and it actually imparts the same shimmery, dreamy quality on the original, even though the effect is also sped back up.
To be fair, if you speed it back up by only the mandated 8x factor, don’t shift the pitch back up and leave the reverb on, it’s actually still not a half bad tune to listen to.
Nonetheless, it’s an interesting thing. In a few tracks well prior to the ones on the site here (say around 1991), I used samples of me speaking that I’d made when I had a cold then both slowed down and pitched down by a factor of around 4 to 6, and they sounded awesome, like dragons. I’ve used small chunks of my own material resampled and stretched or compressed in time before. Still, this makes me really want to take some earlier ramp tracks and mess with them to see what kind of new material I can generate.
October: Concert Month Redux
Aug 17, 2010 Live Music
I know I posted about this last month, but seriously, what’s with October? Every good concert coming up in Toronto seems to be in October. So far I have tickets to six upcoming concerts, and they’re all in October. (I hope nobody I speak to in October is expecting me to be able to hear them.)
Right now my October Insanity Roster looks like this:
Sat., Oct. 2nd: Swans at Lee’s Palace.
Sun., Oct. 3rd: Deerhoof at Lee’s Palace.
Tue., Oct. 12th: Howard Jones at The Mod Club.
Wed., Oct. 20th: Four Tet at The Mod Club.
Sun., Oct. 24th: Ryuichi Sakamoto at the Queen Elizabeth.
Wed., Oct. 27th: Recoil at the Opera House.
As far as I know, the Ryuichi Sakamoto show is the only one that’s seated, so if anybody wants to join me for any of the other shows, it shouldn’t be a problem going together if you just buy tickets on your own.
Still no news on any North American dates for the two-day Einstürzende Neubauten 30th Anniversary tour. 
New Track: Rising, Falling
Aug 9, 2010 tracks
I don’t think this is the first time I’ve used more than one “master” sequencer in tandem, but I think it’s the first time I used three. Parts of this were done in Logic Pro 8, parts in Nodal, and parts on the modular system driven by the STG SoundLabs time modules. A lot of the later steps were purely working with audio, as well, rather than keeping things as MIDI until the last possible moments. A lot of the tracks changed radically after initially being rendered as audio. Usually once I’ve “printed” a track as audio, I meddle with it only very little.
Rising, Falling: (download)
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
As always, I hope that you enjoy it.