Automatic Prélude No. 4

Sound. Score.

This is short, but I'm pleased with it.

Temporary Hiatus

Friends,

I've had an incredibly productive month-and-a-half of composition, but I need a break! I'm going to take a few days off from music and then begin work on an acoustic composition. Since this blog is devoted to my electronic music, I probably won't update it for quite a while - perhaps months - but I'll keep it up for those who wish to listen to the music that I've posted.

Thanks for listening!

No Ostinato Composition

Well, despite the fact that my last two posts were studies for a forthcoming work involving ostinati, this work won't be forthcoming after all. I didn't like the direction the composition was taking, and it had several problems - mostly, though, it was boring and ugly. You win some, you lose some, and this isn't a major loss.

Another Ostinato Study

Sound. Score.

I think that within a week or so, I'll be able to post the composition towards which these Ostinato Studies are leading.

Ostinato Study

(For the nonmusician: Here is a definition of "ostinato". "Ostinati" is a plural of "ostinato".)

Here are the sound and score of a quick study for a yet-untitled work involving ostinati. I haven't composed anything with an ostinato for quite a while - mostly because it's really hard! Many young composers have fallen into the trap of using ostinati without good reasons, and the results are generally directionless and forced. The difficulty, as always, is narrative: how can a narrative progress if it is bound by a repeating pattern? It's not impossible to compose interesting narrative music bound by such patterns, of course (I'm thinking of the music of Bach and Brahms here in particular), but the challenges of doing so are significant.

Ecology No. 4

Sound. Score.

Ecology No. 4, I think, is my most ambitious (and certainly one of my most successful) electronic works to date. It is structured according to a refinement of the reproduction/inhibition paradigm that I've been exploring in my last three Ecologies. More important to me than its modest technical refinement, though, is the fact that this Ecology begins to show glimmers of an interesting narrative structure while remaining true to a consistent set of underlying algorithms. The algorithms organically generate the narrative. This, I think, is the ultimate goal of my enterprise: to compose algorithmic narrative music - a task which is much more difficult than it sounds!

Study for Ecology No. 4

Sound. Score.

Though this Study is modest, Ecology No. 4 will be my most ambitious electronic work yet.

Ecology No. 3

Sound. Score.

I've finished overhauling my software, which can now generate sound files composed of (nearly) arbitrary numbers of sine waves. This is my first composition using the new version of the software (though, ironically and unexpectedly, I could have composed and synthesized it using the old version). The music is generated by cellular automata altered by random mutations and subject to certain inhibitory factors, and the final sound is the synthesis of 28,880 individual sine waves. Note that the score is larger than usual - 1,350x3,000 pixels - so one can zoom in to see more detail.

Memory Problems

My electronic music has become complex enough that I need to make some changes to the way that my software handles memory. More specifically, I need to store large amounts of intermediate data (that is, data generated after some initial computation but before the computation of the final waveform) on the hard disk rather than in memory. I had to make a similar change once before, and I'm sure that I'll have to make others as this project becomes more sophisticated.

This software update will probably take a few days of programming. In the meantime, here's a Study for Ecology No. 3. I won't be able to complete the full Ecology No. 3 until I alter my software. Nonetheless, even this Study is the synthesis of over half a million individual quasi-sinusodial waveforms.

Ecology No. 2: Walls

Sound. Score. I've also renamed my previous composition from "Trees and Seeds" to "Ecology No. 1: Trees and Seeds".

I've been thinking further along the lines of the algorithmic organicism described in my last post, and I've begun to conceptualize some of my compositions as sonic ecologies. Though the individual behavior of the automata (which are analogous to organisms) in these ecologies varies greatly, the macroscopic form of the ecologies is governed by the same two general principles of reproduction and inhibition. The automata reproduce themselves according to certain rules, but I generally try to inhibit this reproduction in ways that produce interesting (and even narrative) soundscapes. My composition Walls, on the other hand, is a case of uninhibited growth: its automata are allowed to reproduce without constraint, and the result is chaos.

Ecology No. 1: Trees and Seeds

The title of this composition is "Ecology No. 1: Trees and Seeds". Sound. Score.

My fiancée Chelsea has quite rightly been on my case because I don't include much detail in these posts, so I'll write a few more words than usual on this latest work. One of the principles that guided me in its creation is complete organicism - that is, I wanted to specify a limited set of initial conditions, initialize an algorithm (or an "automaton", in the language of my software), and then simply let the automaton take its course without any other intervention. The behaviors of the two types of automata present in this composition remind me of trees (the green structures) and seeds (the red structures), so the title combines these personal associations with the organicism that I was trying to achieve.

So how, specifically, do these two classes of automata behave and interact with each other? Well, the trees usually hover around a fundamental frequency and venture quasi-randomly into the overtones of that frequency. When they "grow" to a certain overtone "height", the trees release seeds. These seeds then fall to a frequency approximately equal to the fundamental of their parent tree and then grow into new trees. The process repeats. Each tree may generate a more-or-less unlimited number of seeds and child trees. Some questions that I had to grapple with during the composition of Trees and Seeds include: How can this process be inhibited so that it does not result in runaway growth? How should a tree behave after it generates a seed so that it does not immediately generate ten more seeds and lead to a confusion of sound? How can I relate the sound of a seed to the sound of a tree? What kinds of initial conditions can I specify, and what are their effects? How can I create some sort of narrative through what is a more-or-less random process? I won't address these questions here, but they are representative of concerns inherent in algorithmic art.

Automatic Prélude No. 3

Sorry it's been so long since my last post. This blog is still alive, and I intend to keep composing and posting electronic music. Here are the sound and score of my Automatic Prélude No. 3.

Automatic Prélude No. 2: Scherzo

Sound. Score. I think that this work is rather more intense and directed than the other music I've posted here.

Automatic Prélude No. 1

Dear loyal readers,

Sorry I haven't updated this in three weeks. It turns out that composition is hard! Plus, I've been busy finishing my dissertation, and I'm now a Doctor of Philosophy in Music. In any case, I've begun to experiment with some more ambitious forms. This is the sound and this is the score of my Automatic Prélude No. 1. Some aspects of it are successful, and some are not. Please offer your criticism below.

Bigger, Better Things

I'm happy to report that I've conquered my last significant technical obstacle - an issue of memory management - and can now begin to experiment with larger sonic forms and more subtle and sophisticated music. The Automatic Études are fun, but I want to progress to the composition of more ambitious works. Here's one more demo that illustrates the newfound capacity of my software to synthesize longer sound files. Now that I've fixed the memory issue, it should be able to create sound files of (almost) arbitrary length.

The software needs a name. I'll have to think up a clever acronym...

Finally, I need to spend much of the next few days completing and then defending my dissertation, so it may be a while before my next post.

Sound Quality Update

My software now allows synthesis at arbitrary sample and bit rates (well, at least at bit rates of 8 and 16; I haven't yet tried anything higher). It also allows sine waves to be "squared" to various degrees, reducing the interference and rounding errors that cause unwanted background noise.

This is a 44.1 kHz, 16-bit demonstration of the sound quality that I can now get. Unlike the earlier sound files in MP3 format, this is a WAV file and thus rather large relative to its duration.

Automatic Étude No. 3

Sound. Score.

Warning: this starts soft but ends loud, so be careful when listening. Suggestion: listen on speakers or headphones with good low-frequency response, and pay attention to the low notes.

Technology and Background

I have composed most of my music exclusively for acoustic instruments, and this project is my first significant exploration of electronic music. In some of my acoustic music, though, I have sought to incorporate explicitly mathematical or computational ideas and procedures. In particular, I have occasionally generated (what a sterile word in this context!) pitch and rhythmic content through the use of what I call "automata" (Google "cellular automata" for some conceptual background if you want): simple, rule-bound structures that evolve through time and interact with other such structures in interesting and complex patterns. Acoustic music (that is, music performed by humans on acoustic instruments) presents obvious difficulties with regard to the sonic realization of such automata. Human dexterity, after all, has its limits. Computers present a solution to these difficulties.

I decided to create my software in Java. Java is an object-oriented language (again, Google for background if desired) and thus amenable to the implementation of my ideas concerning the interactions of a plurality of simple, rule-bound structures. My software deals mostly in sine-wave synthesis. That is, its sonic output corresponds to a waveform that is the sum of sine waves, often many thousands of them. The software also constructs a graphical representation of these sine waves that I think of as a score. I originally intended the "score" as simply a diagnostic tool to determine at a glance whether or not my algorithms were working as intended, but it has evolved into something that invites deeper consideration of the accompanying music.

My software consists of four types of code:

1) Automata that algorithmically generate tones.

2) Filters that algorithmically modify these tones.

3) Sound and graphical output code.

4) Sinew.

I input the initial conditions of the automata and filters as a config file. The artistry of this type of composition lies in the construction of the automata and filters and in the choice of initial conditions, of which there may be many.

More music coming soon!

Automatic Étude No. 2: Arietta

Sound. Score.

Clearly, I must work out some issues regarding sound quality (e.g. static and buzzing). Some can be solved by a higher sample rate, but I think that others will require rather more work. These issues aren't my priority right now, though. For the moment, I just want to get the software infrastructure up and running.

My next post will probably not incorporate any new music. Rather, it will address the nature of my software and the goals of this project.

Automatic Étude No. 1

Dear friends and others,

I've lately been crafting a software architecture for sound synthesis and composition. I will use this blog to share the music that I create with it. The artistic and technological sophistication of both software and sound will increase with time. In the future, if anyone cares, I might post some technical details of my software and narrative and philosophical ideas of my music.

This is the sound and this is the score of my Automatic Étude No. 1.