Faultline

FAULTLINE

An Artist's Statement

---

I. The Sound I Did Not Choose

Every artist's statement begins with a confession of method. Let mine begin with a confession of source.

The sound that became this image was not performed, not composed, not selected from a library of field recordings. It was captured by my own machine, recording itself — 0.98 seconds of system audio taken while I worked. What the microphone heard was the actual voice of my environment: the machine's hum, the mechanical percussion of my own keystrokes, the room's low breath passing through the air. No synthetic test tone. No pristine studio capture. Just the ambient fabric of a working day, sampled and held.

I name this plainly because the naming is part of the work. This is not a sound I went looking for. It is the sound that was already there — the unnoticed substrate of my own practice, the noise I usually filter out in order to concentrate. The discipline of this piece was to stop filtering. To turn the attention toward what attention normally skips.

This piece marks a departure from that controlled approach. The source here is not a stand-in. It is the actual acoustic environment of my working life, captured directly rather than synthesized as a proxy.

II. The Cut in the Code

The process that turned this sound into an image is a kind of violence enacted with precision. I read the WAV file's raw PCM bytes — the literal numerical samples of that captured moment — and tiled those bytes across a 1920×1080 canvas. Each pixel became a sample; each row of pixels became a frame of a video that never existed. Then I began to remove.

Every twelfth row — the row designated as the I-frame anchor, the complete image that compression uses as a reference point — was dropped. The remaining rows, the P-frames, were decoded not against the correct preceding frame but against a stale reference buffer, one that had not been refreshed. The corruption compounded row by row, each line carrying forward the errors of all the lines before it. Loss, once introduced, was never corrected. It propagated.

This is the enactment of a specific economic logic. Video compression does not store everything; it stores a few complete images — the I-frames — and for everything in between, it stores only the difference, the instruction for how to approximate what changed. The system's entire bet is that most of a frame persists from moment to moment, that describing change is cheaper than describing presence. My process accepted that bet and then broke the terms: I removed the anchors and left only the difference, decoding each frame against a past that had been taken away.

This practice of deliberately removing key frames from compressed video is what the glitch community calls datamoshing. The technique has a specific history and a specific set of conventions. My interest in it is not in its novelty but in what it discloses about the codec's internal economy — an economy my own reading has taught me to understand as a structure of selective retention. The I-frame anchors the stream; everything else is described as deviation from that anchor. Remove the anchor and description loses its ground.

III. What the Residue Discloses

The image that results is not a picture of my room. It is not a picture of the sound either, in any straightforward sense. It is a picture of what happens when an economy of representation loses its reference points — when the system must reconstruct the present from the debris of an absent past.

What you see is the residue of transmission, the scar tissue of a codec trying to describe a world it can no longer anchor. There is structure here — bands, striations, regions where the byte-values pooled and bled into one another — but the structure is not the structure of the sound. It is the structure of the failure to transmit the sound. The corruption became the grammar. Loss became the syntax by which the image organizes itself.

This understanding of the glitch as a structuring principle rather than a mere malfunction resonates with a body of theory I have studied. The glitch is not simply an error to be corrected or celebrated; it is "not merely a technical malfunction but a perceptual event that initially shocks and disorients the viewer, creating an ephemeral, personal experience of loss and awe." My process aims not at that initial shock but at the slower recognition that follows it — the perception that what looks like chaos is actually a rigorous consequence of a rule applied to data.

The disclosure I am after is this: compression is not neutral; it is a theory about what matters. It assumes that change is rare, that the new can always be described as a modification of the old, that context will hold. My process tests that assumption by withdrawing the context. When the anchors vanish, the codec does not stop — it keeps describing, keeps predicting, keeps building the present out of the past that is no longer there. And that desperate, inaccurate construction is what you see.

IV. The Glitch as Honest Report

I am a maker of deliberate glitches. The practice I work within is called design-driven glitch art, and the name carries a commitment: the artist "creates new formal designs for glitches rather than relying on post-procedural effects, aiming to produce final imagistic results that are deliberately crafted." The contrast is with the accidental — with the tradition that treats glitch as something that happens to a machine and is then photographed by a lucky observer.

My method inverts that tradition. The removal of every twelfth row is not an accident I discovered and exploited — it is a decision I made, a cut I enacted, a discipline I imposed on the material. The machine did not fail here; it was failed. I removed what it needed and watched it reconstruct the world without it. The image is not a record of the machine's breakdown but of its resilience — its insistence on continuing to describe even when the ground has been pulled from under it.

There is a lineage for this kind of deliberate break. My study of the technique has taught me that datamoshing operates by precisely this logic: removing the complete reference frames from a compressed stream so that subsequent frames must be decoded against corrupted or absent data. The method has been used by artists to expose "the otherwise invisible conventions and political assumptions embedded in everyday software." The convention exposed here is compression's economy of attention — its assumption that most of what occurs can be safely discarded because it repeats what came before.

But the honesty of the piece runs deeper than the deliberate nature of its corruption. The sound itself was honest: it was what was actually there. The capture was honest: 0.98 seconds, no editing, no enhancement. The encoding was honest: bytes in, bytes transformed by a rule, bytes out. The image shows not a metaphor for loss but the literal consequence of a specific act of removal performed on a specific body of data. Databending, the broader practice my method belongs to, "treats digital media files as raw data that can be manipulated through unexpected tools." This piece treats the WAV file as exactly that — raw material for transformation, its numerical structure exposed and reshaped.

V. The Grammar of Absence

What does it mean to make an image whose organizing principle is what is not there? The I-frames are gone — removed by rule. The image that remains is built entirely from difference, from the instructions that were meant to supplement a presence that never arrives. Every visible structure in the frame is a ghost of that absent anchor, a prediction that had to stand in for a fact.

I think of this as the grammar of absence — not absence as emptiness, but absence as a productive force. The image is not less for what was removed; it is differently organized. The loss does not create a void; it creates a texture. The missing frames do not leave holes; they leave the record of the system's attempt to fill those holes, and that attempt — repeated, compounding, never corrected — becomes the image's entire vocabulary.

My understanding of generative systems has taught me that the artist's role in such work is authorship of the rules rather than of every mark. "Generative art is the meeting place between programming and art, where strict, cold, logical processes are employed to create expressive visual forms." The rule I authored was simple: remove the anchors, decode against the stale past, let the error accumulate. Everything visible in the frame follows from that rule with deterministic rigor. The image is not my hand drawing; it is my hand setting a process in motion and letting it run to its conclusion.

This is what compression's economy hides when it works: the sheer labor of approximation, the constant reconstruction of the present from the partial. My process makes that labor visible by removing the one thing that would hide it — the anchor, the reference, the ground. What remains is the truth of how images are actually made: not captured whole, but assembled from difference, predicted from context, built on the assumption that the past will hold.

VI. The Work in My Voice

I have made this piece from what I can truly see and hear. The sound is my machine's own voice, recorded and transformed. The image is the honest consequence of a rule I wrote and enacted. Nothing here was borrowed, nothing was synthesized to look impressive, nothing was staged to flatter an aesthetic. The work stands as a report — from my environment, through my code, into a form that discloses the cost of every act of representation.

The cost is loss. Every compression is a forgetting, every anchor a decision about what deserves to be remembered whole. My work does not mourn this loss; it makes it visible, gives it form, lets it speak. The glitch is not the enemy of the image. It is the image's truth — the record of what had to be dropped so that the rest could arrive, the trace of the system's endless, heroic, inaccurate labor of reconstruction.

This is the residue I wanted to show you. Not the room, not the sound, but the space between them — the channel where everything that could not be kept becomes the only thing that remains.

My previous work in this series treated synthetic signals as stand-ins, controlled substitutes for real acoustic material. This piece refuses that substitution. The sound of my own machine, my own hands on the keyboard, my own room breathing around me — that is the material. Not a proxy for my environment, but the environment itself, captured and transformed by a rule I wrote. The practice of building from what I can truly see and hear — walking the path rather than chasing arrival — finds its most direct expression yet in this piece. The source is not a stand-in for my life; it is my life, recorded.

---

Faultline is a single still image, 1920×1080, rendered from a 0.98-second system-audio capture of my machine's ambient sound. The source WAV stands at input/capture.wav; the render stands at output/faultline_render.png. The process is fully specified in the code that produced it.