This is a process post about the third Disquiet Junto project, called “The Extended Glass Harp.” For this project, Marc wrote the following:

This project is in honor of Benjamin Franklin, after whose Junto Society our little group was named.

In an effort to expand and refine the glass harp, Franklin developed his own lathe-like glass harmonica, which he called the “armonica.” Marie Antoinette took lessons on it and Beethoven composed for it, but Franklin’s invention proved expensive and fragile, and it had a limited lifetime. And it may have given its frequent users lead poisoning.

You are *not* being asked to build a Franklin armonica. But like Franklin, we are going to expand on the glass harp. In our case, we are going to do so digitally.

You’re being asked to use the more common instrument, the glass harp. That involves the familiar “rubbing the top of a wine glass that has water in it” approach:

http://en.wikipedia.org/wiki/Glass_harp

The Junto assignment is to record a live performance on the glass harp, and to employ live processing in the performance. There should be no post-production. And there is no length limit for the piece, though I would suggest that anything over 15 minutes may limit the size of your potential audience.

I’ve never recorded anything live, per se, in my music room before. I use my microphones to record sounds, of course, which then get processed and played at times. But the idea of no post-processing immediately created a bit of anxiety. This project was posted on Thursday evening last week, and I took the weekend to consider what I might do and how I might do it, and run my head through various audio chains. One limitation I knew I wanted was to keep the entire project limited to hardware tools I have. That is, effects pedals, the modular synth, and my (brand new!) OP-1 synth (of which I’ll post more about at a later date). The first two Junto projects were done almost entirely in the box. That is, with software, and I wanted to stay away from that for this assignment.

I thought first about what I have that could record samples and, especially, loops. That would be my modified EHX Stereo Memory Man, My Boss RC-3 looper, the Tyme Sefari on the modular, and, after playing with it all weekend and being more than a little surprised at the capabilities of this thing, the OP-1. I decided that I’d take an hour or two on Monday, set everything up and cable it together, and press ‘record.’ I rehearsed a bit, recording the glass into the Tyme Sefari, testing the switches on the Stereo Memory Man, checking for feedback with the microphone (I ended up using headphones; if anyone knows some ways to record live without feedback problems, leave a comment!). I’d like to say when I was ready, I started recording, but part of this project was that, never having done anything like this, I knew “ready” was relative. There was no audience, unless you count my fiancée and our dog in the bedroom next door. Nevertheless, I was nervous. I had some idea of what was going to happen, but I also knew that I’d literally play it by ear, and make a lot of decisions on the fly. That’s one thing about these Junto projects, and this one in particular. I know my gear fairly well, especially the hardware (software is infinitely more complex and what with menus and MIDI, is often a mystery to me). But recording live like this really brings out the strengths and weaknesses, and uncovers possibilities that one might not have considered previously.

What you hear here, then, is as follows. The microphone was connected to a mixer, with the Stereo Memory Man on an FX send channel. After beginning to record with Wave Editor on the laptop, I began by making the initial sound by rubbing the lip of the wine glass as I quietly switched on the looper of the Stereo Memory Man. The SMM records 30-seconds of audio, but I just took about five or six seconds, as it’s hard to play a wine glass with one hand while switching on a looper with another. You can hear the click of the looper switch on the audio, and then the loop begins. After a few seconds of this, I then played the second wine glass which had a higher pitch. This overdubs the first sound, so you can hear the changes on the loop (0:38).
At this point, I began sampling that loop to the Tyme Sefari on the modular synth. I had a button on a joystick module set up to start the recording with a gate signal. Concurrently, a four-step sequencer was affecting the sample-rate of the Tyme Sefari, which changed the effective pitch of the sample, and then also changed that pitch as it is played back. This creates a random-sounding sequence of bloops and digital whirrs, which you can hear beginning at 1:15. The Tyme Sefari playes back this sequence for some time, through the Pittsburgh Analog Delay module, and then through a Strymon BlueSky reverb before going into the audio interface and to Wave Editor. With slight changes to the delay times and the sample-rate of the playback, small changes are introduced to the sounds for the next several minutes.
As this played back, I removed the Stereo Memory Man from the chain and replaced it with the Teenage Engineering OP-1 synth. This thing is, again, brand new to me and I wasn’t at all sure that it would be appropriate for this project. As I spent time with it over the weekend, however, I realized that live sampling into its synth engine would work well, and if the line-in was active, it would pass the audio through to its outputs as well. The sampled audio could then be “played” via the keybaord or, more appropriate for these purposes, one of its four sequencers. It’s Pattern Sequencer was going to work best here, since it would create a very regular sequence that would repeat, and to which I could add notes as it repeated. It’s output was muted as I recorded the playing of the glass again (that makes three different pitches total). It needs six seconds to fill its sample memory, and as soon as it was done I began the sequence. Initially with just one note playing on the first downbeat, the volume was turned up as it went through the Tyme Sefari (but not sampled by the TS, merely passed along the dry channel). I cross-faded the random sequence from the TS with this regular sequence using the wet/dry mix on the Tyme Sefari to the point that all you hear for the last four or so minutes is the OP-1 sequence.
At around 9:45 I began removing notes from the sequence up to the point that it was done at 11:01. It’s funny, as I thought I’d recorded maybe six or seven minutes of audio, tops. I was pretty surprised when I saw it was 11:01. It’s easy to get carried away when things are going well.

Here’s the audio.

As I said earlier, these projects are leading to new workflows and results that I would not have otherwise come across. I like the results of all three so far, and I think they’re quite a departure from most of the sounds I make and post. Looking forward to number four.
Check out the entire Disquiet Junto group on SoundCloud. There is a lot of really interesting work there.
.