For live electronics (supercollider) and 4 channel audio.
This piece is the first of a series (still in progress) of pieces that explore feedback networks, interactivity (perceived) and aural serendipity. The recording below is one realization of the piece as each time the code is executed the results vary. It has also been modified from the original (for four channels) for stereo.
It has been almost five months since MuCo performed for the 2nd time at Ibeam and for one reason or another the recording has sadly sat neglected on a hard drive since then. I finally dug it out, mixed the tracks together (output of kane + scacinto) and here it is for your listening pleasure.
(Just in case you are wondering what the laser-gun sounds are that occur at the beginning and, even more noticeably, during the middle of the recording [11:18 ->], they came from a magical black box that Kane made and I played without knowing how it worked. Hilarity.)
Here is the somewhat-delayed recording of our set on the Subtropics festival in Miami on March 20th. I took some video, but it was terrifically dark and the audio was useless as well. I know Gustavo recorded the audio from the room, but I haven’t gotten my hands on it yet. The recording below was recorded internally (and separately) on our two machines and combined afterward, so no room on this one. If I get a good room recording, I will post it here.
Also, I think there are some good pics from the event that I haven’t gotten yet either, so I’ll post those if and when I get them. Until then, enjoy the show!
(headphones or a system with a subwoofer is highly recommended!)
In anticipation of our upcoming performances March 12th and March 20th, we (MuCo) improvised a set tonight using our new patented MuCo Flouper® and recorded the sonic joy that resulted. Here, for your enjoyment, is the 35 minutes that occurred AS it occurred, recorded in Hi-Def audio and reproduced for you in stunning stereo!
Here are some pictures and a short video from FridayNightThing II (Feb. 25th.) A big thanks to Heather for taking all the pics and video, editing them together, and putting them online. As always, a big thanks to all who came, participated, and spectated. The next FridayNightThing will be on Friday March 25th. Get out your pots and pans and plan on joining us!
A couple of weeks ago I posted some audio and code examples of a software matrix mixer I made with SuperCollider. I thought a post going into some detail about feedback and how to code it may be of help to those new to the concept as a creative tool, or to using it within SuperCollider. Lets first start with some basics of a feedback circuit.
The necessary components of a feedback system (the bare bones) are an input source (some sound, if only line-level hum) and an independent loop with a gain control. Check out the diagram below.
diagram showing basic feedback circuit
The feedback loop can be accomplished in a number of ways. If you have ever used a mic in an auditorium (or been to school for that matter) you have already probably experienced feedback in the form of a high-pitched squeal. The microphone picks up your voice, is routed to a mixer, is amplified (gain) and played out the room speakers. The sound coming out the speakers is then picked up by the mic in addition to your voice. The squeal occurs when the amplitude of the input signal (your voice plus the amplified version coming back over the speakers) is loud enough that with each cycle the overall gain increases. Cut the amplification out to the speakers and the squeal stops. The quality of the feedback (the frequencies amplified) have to do with the microphone’s resonant frequencies in combination with a lot of other factors (room size and resonance, quality of the audio gear being used, whether the signal is balanced or unbalanced, matching or mismatched impedance, etc.)
In the mixer feedback post I did a couple weeks back, the input signal was the inherent noise of the analog mixer with gain applied at each stage of the 4-stage feedback loop. In any analog system there is always noise present because of the nature of electrical circuits and the power required to make them work. The same is not true for purely digital (software) feedback systems so some generated sound input signal is necessary.
Below is some very simple SuperCollider codez to demonstrate this simplest of feedback loops and a recording so you don’t have to run the code. If you do run the code in SC, be careful when moving the mouse to the right of the screen. As you approach 1, the feedback loop will become increasingly loud until at 1 and beyond, it becomes exponentially loud until things break. (And by ‘things’ I mean your speakers and/or eardrums and/or the SC server :P) In the recording below, i just move the mouse to the right to “max out” the feedback loop, then drop back after it maxes out.
Here is another version with a simple half-second delay built into the processing section. Note that this rids us of the ear-piercing noise that occurred in the above example. In the recording I’m moving the mouse to the right, then back and forth across the screen to make a more interesting texture. I then just let it build up with the mouse all the way to the right.
The next two examples are of slightly more interesting feedback circuits. In the first version I have fixed the gain at 1.1 so it the sound eventually reaches saturation and does not return. I’ve removed the limiter and replaced it ‘.clip’ which squares off the wave form resulting in audible distortion of the signal. I have also added a random line generator ‘speed’ to control some aspects of the sound, the resonant frequency of a low pass filter and the delay time.
While these samples are a few steps away from being art, they show the power of possibility lurking in feedback circuits. The truth of synthesis is that, with a few exceptions, the more intricate (complex) a sound is, the more realistic and better it sounds. Adding a feedback loop to a sound introduces a few more layers of complexity and can have beautiful as well as destructive results.
If you are interested in hearing some truly excellent feedback music, check out the work of David Tudor. The complexity of his circuits is both astounding and elegant, and the sonic results are really fantastic.
Following any of the steps below to create feedback loops with mixers can harm your gear and more detrimentally, your ears. The results are often unpredictable and almost always extremely loud. The pulse waves created by these kinds of setups and heard in the recording below are very hard on the ear mechanism (as you will be able to tell by listening.) Please take all precautions to limit the amplitude of your speakers and, if listening on headphones, to start with the volume very low and turn it up as needed. If you plan to attempt the following setup or one like it, start with all volumes at the minimum and raise them once you know what your results are going to be.
Note: the piece begins very quietly, the first loud sound is around 1:26.
The following is a list of equipment used in the above improvisation.
Dell Latitude D620 (1.6gHz, 1GB RAM) running the latest PureDyne distribution
Kane recently played a few recordings for me of experiments he had done with feedback systems created using his 1204 mixer. The sounds were appealing and I thought it would be fun to see what it was like to make music with only a mixer for an instrument. My 1204FX has on-board DSP that Kane’s model does not. Normally, I do not use the processor at all, but for this exercise it was useful in adding variation to the signal flow and achieving a variety of sonic results.
Last night I experimented for about 2 hours with different routing schemes and to get used to controlling the mixer as a sound-generator. I recorded 8-10 tests and ended up with about 45 minutes of pretty good material which I may use at some point in the future. I then recorded 1204-10-29 in one take, using only the 2-channel output from the mixer. There is no additional material in the recording, nor any post-processing aside from normalization. The following is the routing recipe I used.
Routing the 1204FX
The first pair of feedback loops was connected as follows:
Alt 3 output –> channel 1 input (trim at +60) –> sent to Alt 3-4
Alt 4 output –> channel 2 input (trim at +60) –> sent to Alt 3-4
The second pair of loops was connected like so:
Aux Send 2 –> channel 5/6 L (+4) –> Main Mix (no Alt 3-4) –> Aux Sends 1-2 alternately as desired
Aux Send 1 –> channel 7/8 R (+4) –> Main Mix (no Alt 3-4) –> Aux Sends 1-2 alternately as desired
Aux Sends 1-2 at +15
Aux Returns at +5 to +10
Aux Return 1 to Aux Send 1 at +5
The reverberation heard is the built-in “Chapel” reverberation, program 19 on the mixer. I used the Control Room R & L output channels to route the audio to my laptop for recording. I monitored the sound using the headphone jack on the mixer with the volume as near to zero as I could get it. (At some points this was not enough and I had to quickly pull the phones off.)
Useful parameters for making music
There are many ways to achieve sonic variation within the mixer. The controls I used were the “pre” buttons for each channel, which control signal flow to the main mix and the aux sends, the faders for each channel plus the ALT 3-4 and Main Mix stereo faders, the “ALT 3-4″ buttons, the AUX 1-2 faders, the pan controls, and the 3-band EQ for each channel. (Is that everything, you say? Almost, I didn’t touch the trims, the low cuts, or the aux send knobs below the DSP area.) The controls I used the most were the volume faders and the 3-band EQs. All of the frequency variation (thumping lows to screaming frequencies around 12k) was accomplished by turning down two of the three EQ bands, and playing with the remaining band while simultaneously working the volume fader for that channel.
If you are interested in experimenting with a mixer like this, trial and error will be your best guide. Try making the channel settings similar for all channels and then changing them one by one to clearly hear the results. Or try using only 1 or 2 of the channels and later adding the rest one by one. Most of all, play with the levels a lot: I noticed that in several instances minute changes to a single channel produced startling results. Also get to know your routing: changing the ALT 3-4 stereo faders will affect all of the channels using the ALT 3-4 pair, while playing with the gain of an individual channel will only affect other channels that share its signal path. By bypassing the aux sends (the DSP) you can have two layers of sound, one processed and the other dry (you can hear this clearly in my piece), so experiment with foreground and background layers.
Here, again, for your edification is my improvisation… I know you don’t want to scroll all the way back to the top of the page.