thesis update

It's been quiet on the home front here, but the end is drawing near (2 more weeks before the final presentation on 5/7, to be exact), and there is quite a lot to update.

This project has changed a lot since its conception, but many new ideas and potential future project concepts have sprung up along the way, and I am pretty pleased with how the project at hand is shaping up.

The original plan was to embed circuitry in felt itself... something I am still interested in exploring. I wanted to design felted audio signal processing circuits that would accept an analog signal in and output it back out after running through a series of filters.

I am near completing a similar project, only I decided to begin the research from a simpler place, working my way up the ladder of experimentation.

Having said that, felted signal processing has become a set of analog signal processing circuits consisting of a chorus effect, delay effect, tremolo effect, and finally, distortion effect. The processing circuits live on pcbs instead of within the felt itself, reassigning the felt's role from circuit board to soft potentiometer. The felt is directly integrated into the hardware, allowing the user to tranform various stages of the processing circuits as the audio signal passes through the mix of natural wool and steel wool.

For example, with the tremolo circuit, there are two felted objects. One controls the rate of the tremolo, while the other controls the depth. There are also two hardware pots available for setting the initial rate/depth. The felted pots offer expressive control over base settings that the user can dial in with the hard pots, for the purpose of introducing some constants to the system. The hardware pots can also be turned all the way down, leaving the signal processing to as much instability and momentary control as desired by the user. The goal is to set a fixed range over which the felt can distort the sound.

Each effect is daisy chained to itself and may be turned on or off within the chain by a switch. However, after I have all the effects and felted controllers finished, I plan on building a 3x3 resistor mixing matrix, that will provide multiple opportunities for feedback, amplification, and attenuation of the signal. Instead of each effect simply being daisy-chained to one another, they will be tied to various nodes in a 9-celled square, each node have 2 or 4 possible routes it can travel down. It opens the the doors up to all kinds of silliness, for sure.

feedback/delay katzen

original snd:
mew, mew!

and fubar'd sound (delay over two channels with feedback):
ch 1: 2ms, ch2: 50ms
both ch: 2ms
both ch: 10ms
ch1: 110ms, ch2: 100ms
variation, ch1: 110ms, ch2: 100ms
both ch: 125ms
variation, both ch: 125ms
ch1: 240ms, ch2: 250ms
ch1: 300ms, ch2: 250ms
both ch: 500ms
both ch: 1500ms
ch1: 1750ms, ch2: 3500ms
both ch: 4000ms
both ch: 5000ms

snd particles

my max patch:


individual 3 millisecond sound bits at varying frequencies:
50Hz
220Hz
1000Hz
4000Hz
10000Hz

sequences of sound bits at various frequencies and playing for various durations:
5 and 10 millisecond
1 millisecond
3 millisecond
5 millisecond
10 millisecond

feltedsignalprocessing.com

I just created a new space for the home of my felted audio filters. It's in development as you read, but for now I have a little page with a little description holding down the fort:

http://www.feltedsignalprocessing.com (in a new window)

img noise

becomes

this noise from one pass through the img

and

after 10 passes through the img

This is a very stripped down image analysis. I'm just converting the rgb value of each pixel into either a -1, 0, or 1, and then sending it into a buffer. I have a phasor object that takes a second to tween from one sample to the next, hence the rhythmic nature of the noise. These two mp3s are only playing blue and green values.

it ended up sounding like a dot matrix printer... more experiments to come.

Strange Attractors

in S-T-E-R-E-O


Refreshing every 5ms
Refreshing every 50ms
Refreshing every 250ms
Refreshing every 250ms again

Modul8

Some samples of different kinds of modulation i was trying out:

Amplitude Modulation [AM]
bass screenshot
AM bass
AM bass 2

tremolo screenshot
AM tremolo

synth screenshot
AM synth

synth2 screenshot
AM synth2

Ring Modulation [RM]
dialtone screenshot
RM dialtone

poleposition screenshot
RM pole position!

Frequency Modulation [FM]
feedback screenshot
FM feedback with 1 oscillator

jetsons screenshot
FM jetsons

orb screenshot
FM orb

psb screenshot
FM psb

FM harmonix

hisynth screenshot
FM hisynth

pulse screenshot
FM pulse

808 screenshot
FM 808

CUT N PASTE SND

i cut and pasted and moved around txt in .wav files with the help of a text editor. i even transplated parts of some sound files into other sound files. scandalous!!

original Eat
Eat 1
Eat 2

original Latch
Latch 1

original Purr
Purr 1
Purr 2
Purr 3

original Wat
Wat (meow mix)
Wat 1
Wat 2
Wat 3
Wat 4
Wat 6

Sine Waves: catch-up!

Here are a few sound files generated from driving different wave shapes through an oscillator I meant to post a couple weeks ago!

sine wave ramp-up
rectangular shaped wave ramp-up
saw tooth wave ramp-up
trapezoidal wave ramp-up
triangle wave ramp-up

Intonations

Here's a few scales I put together, divided by equal temperament and just intonation:

Equal Temperament
2-note scale starting at 100Hz
4-note scale at 100Hz
5-note scale at 100Hz

Just Intonation
12-note scale, low D
12-note scale, low E
12-note scale, low F#
12-note scale, middle Ab
12-note scale, high C#

additive synthesis

combinatorial sine wavery trickery via max/msp:

50 Hz of BASS
7 harmonic partials of more bass pt I
7 harmonic partials of more bass pt II
mix of sine waves
swell of sine waves
triangle burst
square burst
bouncing sine waves pt I
bouncing sine waves pt II
a horribly synthesized jazz piano.. maybe..

Curated listening

I have made a few selections of sound/music that I found inspiring for various reasons.

The first is a song by the band Bowery Electric, called Out of Phase. They play with texture primarily and stretch the musical experience out in a droney, luscious fashion.

The second track is by a band from the UK, called Broadcast. The song title is Minus Two. Broadcast are well known for their ability to manipulate sounds into all kinds of textures and melodic strings.

Next is a recording I made myself of an escalator railing at the subway station in the middle of Times Square. It's a virtual symphony of mechanics.

And finally! There are a couple of selections I'd like to play from studies done by Mark Whittle, a professor at the University of Virginia. Analyzing Cosmic Microwave Background (CMB) radiation data, he was able to sonify the first 100,000 million years of the universe after the Big Bang (which, according to him, was actually the Big Flash, as the beginning of the universe started with silence, rising to a dull scream as matter pushed out ever further.. pretty neat!). This BBC article has some concise reporting on it.

Listen out / Listen down

While practicing listening downwards into the inner nucleus of the buzz of hallway lights outside of my apartment door, I found myself distracted by the outer sounds I kept hearing. Therefore, before focusing on what was going on at deeper levels of fluorescent lighting, I let myself focus outwardly.

Starting from the light I was standing under and moving outwards, I became aware of the fact that once I stopped to really listen, I could hear everything from the clanging of dishes in apartment 3A to (unfortunately) my neighbor relieving himself over in 4B. Beyond this, cars were easily audible from the street out front, with a couple dog yaps and childrens' squeals piercing their way over from the building on the other side of the road. Moving further out still, I could trace the sound of more traffic rushing down the streets that criss-cross through the East Village, out towards the Williamsburg Bridge. Moving upwards, you can also detect the sound of wind swaying winter bare branches, to the flap of someone's drying rug outside of some unit above, all the way up to the dull scream of a jet slipping across the sky to I can't say where. Actually, the plane felt almost as loud as the cars cruising by out front of my building.

Having got that out of the way, it was easier to come back to think about the lowly light whose buzz nonetheless always caught my attention. It's not just these particular lights in my hallway; it's fluorescent lights with their buzz in every hallway that I've ever been in that catch my attention, specifically when I'm sitting in the stairwell trying to focus on anything to keep myself from being sick, usually around 4am after a party has gone on too long. Listening long enough, you can pick out the different layers of sound, overlapping each other like scratchy gauze. The different frequencies even seem to move in and out of phase with one another, creating a hypnotic lull of monotonous shifting audible vibrations.

For reference, here's a recording of the lights just outside my door.

Loom: Update 6

Okay, I'm still waiting for a space to set my loom up in. And my cat got sick on my laptop, thereby killing it (it's not salvageable apparently) and cutting my access to my max/msp and arduino files thusfar developed. Not a great week.

But I did make some progress, despite that. I re-did my warp, as it wasn't long enough the first time I made it, and it went A LOT faster the second time around (about 2 hours vs, uh, 10 hours last time). So that is now finally complete and attached to my top beam. Other things completed:

- Got the correct bolts in my top beam supports. It's all bolted up to the uprights and ready to go.
_ Got my shed stick cut and bolted to the bottom of the loom (for separating the front and back threads from each other)
- Put together xbee/accelerometer circuits to go on five of the 24 weights hanging from the warp
- Got my shuttle put together with the wireless accelerometer sensor + a lithium battery
- Rewrote my max patch to where it was before

Oh yes, the xbee is an addition, heh. I am combining this final with my Mesh Networking final, and making the weights wireless devices that send their values remotely. I will be keeping a wired version of the code and board handy though, in case some kind of wireless mishap takes place, which seems to always be a danger, and I don't want any of that at showtime.

This next week is just wrapping everything up I guess and well, getting the xbee stuff all working flawlessly as well as tweaking the sound still.

Loom: Update 1

So I've broken the task of building this loom down into a few categories of attack:

Material (yarn)
Loom (building + augmentation of...)
Hardware (sensors, circuits)
Software (music mapping and programming)
Performance

This week I took on a few of the tasks under the Software header. I mapped out how I would use the loom for actual sound production and worked on some test patches in max/msp.

I also decided on what kinds of sounds I am going to produce. I'm going with deep synthy sounds, contrasted with fuzzy, crackly, scratchy sounds. Although it's nothing I'm completely intentionally doing, I'm well aware that many of the sounds I'm going with reference wool, hair, and other organic fibers.

As far as mapping goes, I'm tentatively going with this plan: I'd like the warp (vertical strands) to alter the sounds I've already laid on the canvas (which will be controlled by the weft, or the horizontal strands). The shuttle will either operate as a beat keeper, or it will act as some kind of interrupt as it slides across the warp. Actually, it will become percussive by default, as it affects the warp threads. I'd also like to have panning mapped literally going from left to right across the loom.

This next week, I want to continue working on my patches. The max/msp part feels like the largest challenge so far, since it is the part I'm the least familiar with. I will also try to build out a 'test loom', which will just be warp and weft threads hand-laid out so I can try to mock up what I ultimately want to do with the final loom.

The biggest challenge is working with max/msp. I'm fairly new to this program, and it can seem like a beast, and therefore hard to get started with. I'm overwhelmed by what I can potentially do, and by deciding what I exactly need to do. For now, I'm just accepting midi-data in from a test circuit, and producing sounds with that data.

Quick pcomp sketch

So I was asked to make a little device that addresses a problem on the ITP floor. Well, 'problem' is a very subjective term, and it turned out that my 'problem' with the ITP floor wasn't as much of one for anyone else, thereby pointing out that I might be a little OCD. oh well.

The problem I chose to tackle was the one of the pcomp lab's tool box drawers not always being fully shut (photo coming soon). It's not a *huge* problem to me, but an issue nonetheless, however petty it might be. And anyways, it's the one I chose to solve, so here you go:


doorbell
Originally uploaded by 0x000000

This device is just a photocell + a doorbell buzzer. When the arduino reads a value off the photocell of less than 200 for more than 10 seconds (ideally, this would be set to a couple minutes, so people have time to open the drawer and shuffle about for their item), a very loud, obnoxious, annoying buzz goes off, until the drawer is fully shut again.

arduino src

Artificiel

Artificiel is a group from Montreal I discovered recently (thanks Mouna!) who puts on some pretty awesome shows using objects like rubik's cubes and light bulbs.

I like how the focus is on the performance objects, versus on the humans manipulating the objects. This is how I think I'd be most comfortable performing, as I'm not much of a ham for the stage. Or for anything else, I don't think. Comfort aside, I'd love to be able to pull off such an elegant and engaging performance.

Check out the Cubing performance:

Playing With Food

My final project for physical computing turned a fork and knife into musical instruments. Sine waves are generated and through max/msp, when you cut into different kinds of food on your table. Full documentation can be found here.