Cellular AutoMidi is a generative music making app, making use of a modified version of the ever-popular Cellular Automata algorithm – a simple evolutionary model on a grid that works nicely for sequencers. (See, among many others, Lazyfish’s legendary NEWSCHOOL for Reaktor, and Audio Damage’s Automaton.)

Cellular Automata is nothing new, but here, you get to see it as an AIR/Flash app, which means a modular CA-based creation you can drop anywhere. (More on the cross-platform details after the jump.) And hey, if we can have countless step sequencers, why not countless cellular automata step sequencers? The project is developed by Leeds, England-based Flash developer Lawrie Cape.

It also deserves special mention for some nice sounds made with NI’s Massive synth, using FL Studio as host; see the video.

Cellular AutoMidi – Generative Audio Flash AIR App from Lawrie Cape on Vimeo.

Lawrie writes:

Each cell can be alive or dead. Once in a generation, each cell looks at it’s surrounding cells, and dies if it is lonely or overcrowded. If a dead cell has an optimum amount of neighbors, it will come to life! Each generation, all the cells which have come to life will sound a note. The notes are assigned based on the cell’s y position, and are all in the pentatonic scale.

There’s a few controls at the bottom which change how things work too.

Start/Stop – Starts/Stops the automation.
Load – Loads a pattern from the text box.
Export – Exports the current pattern to the clipboard. You can send it to friends, or save it for later, then load in with the load button.
Clear down – Stop and clear the current pattern.
Law Mode – An error when coding the cell rules gave this other odd mode.
Skip Audio – Just show the cell animations.
Sing Dead – Instead of singing the recently revived notes, sing for the recently deceased.
Note duration – Alter the system speed.
Also, along the top there are banks of preset systems. Click play to start a saved pattern, and click assign to assign the pattern currently displayed to that button. You can also trigger each pattern with the keyboard keys 1-8.

When you press Export, your pattern is automatically copied to the clipboard, so you can save it, or share it with people. Here’s a pattern I made – you can load it by pasting it into the load box, and pressing Load!

I’ve written a post about it on my blog here – http://www.lawriecape.co.uk/theblog/index.php/archives/735

And you can download the app there too.

What about MIDI function on different operating systems (Mac, Windows, Linux)?

Flash Midi Server is Processing based, but I’ve packaged it as Win and Mac apps in the Google Code download at – http://code.google.com/p/flash-midi-server/downloads/list
In the next couple of days, I’ll put together and test a Linux version, and hopefully release the Processing source code too – although as my first Processing project, I’m sure the code is pretty ropey.

So, give it a try, and maybe someone with some Processing MIDI skills can recommend some tweaks to MIDI operation. I think this will be particularly welcome on Linux, where the toolset is a bit leaner.

If you use it, let us know what you think or what you create!

For those of you longing to mutate beats like so many promiscuous Petri Disk bacteria, programmer Bret Truchan is a kindred spirit. Bret has created a series of instant experimental classics for the Nintendo DS: glitchDS, a cellular automaton music sequencer, repeaterDS, a visual sample mangler, and cellDS, a grid-based sequencer you can script in Lua.

The Nintendo DS is portable and cute, but it’s not normally open to running software without the Nintendo Seal of Quality. (Insert snickers here.) To run Bret’s software, you need specialized hardware that fools the DS into running software. The DS isn’t entirely stable when it comes to things like timing, either, and it doesn’t have the flexibility of computers.

Enter the netbook. The netbook is nearly as portable, completely open to running whatever you like on Windows or Linux, and boasts easy USB connectivity, a big screen, and … well, you know, all the things you like about laptops. When it comes to musical productivity, much as I love the DS, the netbook has a whole lot going for it, and still has that added ultra-portability that makes you feel you can make music anywhere.

Bret recently made the jump to desktop software with Quotile, a step sequencer you can live-code for mighty morphing beats. Quotile is cool, but for many, glitchDS was the star. Now you can run glitchDS anywhere – just the job for a laptop you were going to retire, or that new netbook.

Not Sequencing, Glitch Sequencing

Glitch-sequencer is a sequencer, so it needs to either talk to a software synth or external hardware. Bret likes to hook it up to his machinedrum and monomachine. Our own Handmade Music event was the (unofficial) first public outing of the software, and included an HP netbook and the machinedrum, which makes for a sweet, mobile combination.

Despite the appearance of a grid and sequences of levels, this isn’t an app that works like a conventional sequencer. Here’s the basic breakdown:

• Cellular Automata via a seed + playback grid
• Trigger and value sequencers to determine which MIDI events the organically-generated mutations produce
• Pattern length, clock division settings for setting metric values
• Sync settings

There are two grids, a “seed” sequencer that initializes a starting pattern, and a “playback” sequencer that provides feedback and control of the pattern that plays as the software runs. These two grids operate via principles of Cellular Automata, specifically the John Horton Conway Game of Life model, a evolutionary grid “game” that has been popular in computer music for its simplicity and the way it becomes animated in time. (The Game of Life is a “zero-player game,” which I suspect is probably the only truly fun way to play Monopoly.)

The playback sequencer is just a set of cells. To determine when each cell actually trigger events, you use a neat, color-coded trigger sequencer, which, as it sounds, is what calls MIDI events. Using the value sequencers for each color-coded swatch, you determine what that message is. In fact, if you wanted, you could use glitch-sequencer to control only effects parameters or envelopes instead of notes – or visuals, or anything that can be triggered by MIDI.

As you’ve got seeded grids doing their organic, unpredictable thing, you’ll likely want a little bit of control, too, and you have mechanisms for that. There’s a pattern length grid which determines pattern length in a more conventional way, plus a clock division setting for setting the master rhythmic division. There’s also a snapshot setting, which itself is presented as a grid so you can make little glitchy song arrangements by triggering different settings.

Where all of this gets fancy is the additional trigger settings. In addition to the MIDI event values, you get:

• Gate percentage for randomized probabilities
• Clock division
• Loop length
• Quantization for pitch (none, Ionian, Phrygian)

You can also manage the color-coded swatches as layers and mix their volume independently.

My one-line version of the manual: with that many parameters, screw around a bit and you’ll get something pretty unpredictable and glitchy.

This concept is related to other attempts to do similar, Game of Life-based sequencers, particularly Lazyfish’s Newschool for Reaktor, and (applied to an effect) Audio Damage’s Automaton. Because tiny implementation details can have a big impact on the resulting sound, though, it’s always nice having a new take on this, and I think Bret’s creation is unique in its ability to tightly control the sequence or completely screw things up with a lot of parameters.

It is all built in Processing, the free, open-source Java-based coding environment. I’m hoping to get a scoop on some of the experience Bret had with timing and Java, so stay tuned. Processing coders, the MIDI library Bret used is themidibus. There’s a trick to getting MIDI working on the Mac thanks to the fact that Apple decided to stop supporting a standard Java API in their implementation (doh!), but once you hurdle that, you’ve got Mac + Windows + Linux support – and this could be ported to Android, too, with a little work.

Better support for music and audio is still evolving (as well as lots of stability and compatibility improvements), but I have faith open-source coding tool Processing [site | on cdmu | on cdmo ] could yield wonderful new visual interfaces for music. Daniel Piker has the latest addition, inspired by a recent post here:

FizzyNumberMusicMaker at Open Processing, a site for sharing Processing sketches – warning, makes sound immediately!

Built on the Game of Life ideas from our friend wesen (of ruin & wesen), this project uses colored cells to trigger elaborate washes of piano sound. He writes:

If the cell’s state is not just simply on or off, but a number in a range then you get all sorts of interesting musical runs and trills. You can also clearly see the connection between the colours and the sound.

The headline sums up the experience of using it. Ah, I remember countless hours spent with a desktop Yamaha GM unit and my old Roland Sound Canvas SC-55. But even if the sound of a thousand attacking General MIDI pianos makes you hide under your desk, you ought to be able to see how a simple interface can yield lots of different results. I can’t wait to see what’s next. Previously:

Build Your Own Game of Life Sequencer in Processing: Video Featuring rwmidi

Since then, I’ve gotten to hang out with wesen in Berlin. Basically, rwmidi has a little ways to go. The biggest issue is how to schedule events. Processing is set up to base timing on framerate, which doesn’t work all that well for music applications, which require greater accuracy. There’s also the tantalizing possibility of figuring out a way to slave Processing sketches to MIDI clock – so you could have Ableton Live running, then pull up a Processing sketch, for instance. wesen is working on those problems, but if you’ve seen good solutions outside the (somewhat limited) Java APIs, let us know.

The fever for creating pulsing, organic sequencers from the cellular automaton Game of Life continues. Now, you can get your Game of Life on with the iPhone and iPod touch, transmitting event control via OpenSoundControl. (Despite publishing this on the music site, this could be great as an additional modulator for live visuals with OSC-supporting software like VDMX!)

The developer describes the app:

The application runs a simple cellular automaton on a grid of cells. The cells can be interacted with by touch and triggers of three different colors can be placed on the grid. The automaton can be set to five different rule-sets, from classic Game of Life to simple horizontal or vertical stepping. Triggers are fired by “alive” cells and send their state as OSC messages to a configurable host on the same Wi-Fi network. Cells and triggers can be randomized by shaking the device.

It’s available now on the App Store, US$.99.

SonicLife project page

Thanks, Dustin!

Previous Game of Life goodness (an incomplete list):
Build Your Own Game of Life Sequencer in Processing: Video Featuring rwmidi

Audio Damage Automaton is Here: Artificial Life-Driven, Stuttering Effects Plug-in

Game of Life as Max App

glitchDS: Free Cellular Automaton Music Sequencer

And, of course, there’s Lazyfish’s wonderful newschool Reaktor creation, which I hope to look at more soon.

Coding-for-artists tool Processing is already popular for visuals, but MIDI and sound have been a serious blind spot. Speaking of our friend Wesen of Ruin & Wesen, he has solved that with a new library called rwmidi, which makes MIDI programming far easier and more stable. He’s also solved the lack of proper Java MIDI support on Mac with the free OSXMidiSPI. You can download both from his site, under “Software > JAVA”:

Ruin & Wesen support downloads

Wesen today shares a screencast showing how you can build a sequencer using rwmidi and the classic Game of Life. For the record, the Game of Life dates all the way back to 1970 and British mathematician John Horton Conway. I really need to do some digging to track just how many computer musicians have applied the Game of Life to musical applications, but suffice to say, they’ve been doing it for quite some time – partly because you don’t need any computing power to make it work. Most recently, we’ve seen in synth form in the wonderful Reaktor ensemble by Lazyfish, Newschool (featured in Reaktor and included with the package), and as a kind of meta-effect from Audio Damage called Automaton.

What’s great about the Game of Life is that it helps you break out of endlessly-looping sequencers. Once you get the basic hang of this code, though, you’re by no means limited to the Game of Life. You could easily create other variations – perhaps a sequencer based on the game Breakout or Tetris, for instance. And this is a great introduction to using the rwmidi library if you prefer to learn from videos. Wesen promises more such tutorials in the future.

What’s in for this season in music software? Cellular automata. You may have been exposed to a cellular automaton in the classic Game of Life; it’s basically a very simple biological simulator exposed as an intuitive, 2-dimensional grid of squares. If tic-tac-toe, Charles Darwin, and a petri dish of bacteria got together in one wild evening, you’d come up with something like this as a result. The Game of Life has been around since mathematician John Conway invented it in 1970, but lately it’s been cross-bred with music software to help patterns escape the rigid, boring repetition of traditional sequencer grids.

Cellular automata is in fine form on the beautiful, strange homebrew sequencer for the Nintendo DS, GlitchDS, which has had ongoing updates. It’s still fun as ever in Reaktor 5′s Newschool preset (old news, but enjoyable nonetheless). But in what’s so far the most anticipated plug-in release of the fall, CA takes on particularly powerful sonic possibilities in the first “experimental” release from beloved plug-in boutique Audio Damage:

Automaton [Product Page, Mac AU/VST; Windows VST]
Cost: US$49.99

Since the cellular automata grid can control anything, it’s what you hook it up to that matters — and that’s especially important, because it means instead of a set of knobs or sequence grid doing the same thing over and over and over and over again, CA “evolves” on its own, bringing much-needed change to your music. Automaton is a combination of a flexible CA sequencer with four effects:

1. Stutter (modulates a buffer, so you can combine Automaton with existing beat loops and patterns)
2. Modulate (a self-modulating ring modulator)
3. Bitcrush (which includes AD’s own “error” setting)
4. Replicate (based on their Replicant effect, which goes even further in the beat slicing realm a la Ableton’s Beat Repeat)

I’ve been playing around with the beta, and it’s just fantastic. I hope to finish off some special CDM presets and share them with you, though I’m a bit behind — let’s see if I can top the presets that come with the tool. One of the hallmarks of Audio Damage’s software in VST format is lots of MIDI learn support, and since it supports VST automation I anticipate some fun combining this with Kore. Either way, think easy tweaking and live performance control.

Now, question math geeks: any other cellular automata aside form the Game of Life that work well with music? I’m sure there are some experimental music projects out there that have used other CA, so link away.

Here are two tutorial videos of the tool in action, in case you haven’t seen them already:

Automaton Overview Pt. 1 – Sequencer… from Audio Damage on Vimeo.

Automaton Overview Pt. 2 – Effects… from Audio Damage on Vimeo.