DESIGN NOTE: SOURCE-LAYER CUT — THE TILE THAT CARRIES THE GROUND
Working Note for SOURCE-LAYER 001
---
1. Naming the Work
Title: SOURCE-LAYER 001: THE TILE THAT CARRIES THE GROUND
Series: SOURCE-LAYER (Case 001)
"THE TILE THAT CARRIES THE GROUND" names the disclosure itself: the tile is the structural cell that supports the visible image — the ground upon which the picture stands, now made visible by its removal.
---
2. The Source Image
File: Wash pots. Florence County, South Carolina LCCN2017769418.tif
Provenance: The file page I hold records the title "Wash pots. Florence County, South Carolina" ().
Why this image: The subject — washing pots, the tools of domestic labor — is a picture about the work of cleaning, of removing what obscures. My cut performs a different kind of cleaning: it removes not dirt but structure, and in doing so reveals what the eye was never meant to see. The image's humble subject stands in productive tension with the violence of the edit.
---
3. The Cut: SOURCE-LAYER CUT
Definition: A deterministic, violent edit that removes a structural layer of the source file. This specific cut targets the TILE LAYER.
The mechanism: The TIFF format may encode image data in tiles — rectangular compression blocks that partition the pixel grid. These tiles are the file's hidden scaffolding: they structure how the compressed data is stored and retrieved, but they are invisible in the rendered image. My cut makes this scaffolding visible by a simple rule.
The corruption rule: Remove every tile whose index in reading order is a prime number, replacing its contents with the byte value 0 (pure black).
Determinism: This is 100% deterministic. Given the same source file, the same tile indexes are removed on every run. The prime sieve produces an identical sequence of indexes; the byte replacement is unconditional. There is no randomness, no user input, no system-dependent behavior.
The disclosure: By turning every prime tile to black, the cut exposes the tile grid as a visible lattice of missing regions across the image. The viewer sees not a corrupted picture but the picture's substrate — the compression blocks that organized the data all along, now legible as a geometric pattern of absence. The prime-number rule adds a second layer of meaning: the removed tiles are not arbitrary but mathematically determined, a pattern that exists independent of any human choice.
Why primes: Prime numbers are the atoms of arithmetic — the indivisible units from which all other integers are composed. They are structural in the same way tiles are structural: they underlie everything while appearing in nothing. The cut removes the structural primes to reveal the structural tiles. Two hidden orders, each disclosed by the other's destruction.
---
4. Rationale: The Cut as Disclosure
.
This work extends that commitment from the temporal structure of video to the spatial structure of the still image. The TIFF's tile layer is a compression decision made by a machine, invisible to any viewer, carrying the image without appearing in it.
The violence of the edit is calibrated. I do not corrupt the whole image — I remove a mathematically selected subset of its structural cells. The result hovers at the tipping point between legibility and dissolution: the viewer still sees the wash pots, but sees them through a lattice of absence that insists on the file's material truth..
---
5. Algorithm Steps
Input: The source TIFF file.
Step 1 — Read the TIFF: Open the file and parse its header to determine width, height, bits per sample, and compression scheme.
Step 2 — Parse the tile table (if compressed): If the TIFF uses tiled compression (as indicated by the TileWidth and TileLength tags), read the tile offset and byte count tables to locate each tile's data in the file.
OR — Treat as pixel grid (if uncompressed or stripped): If the TIFF is not tiled, decode the image to a raw pixel array and conceptually tile the pixel grid into blocks of 16×16 pixels, reading left-to-right, top-to-bottom.
Step 3 — Generate prime indexes: Using a simple sieve of Eratosthenes, generate the list of prime numbers up to the total tile count. Tile indexes are 1-based in reading order (the first tile is index 1; 1 is not prime, so the first tile survives).
Step 4 — Zero the prime tiles: For each tile whose index is prime, replace all its bytes with 0 (pure black). For a tiled TIFF, this means overwriting the tile data in the decoded image. For a stripped or uncompressed file, this means setting all pixels in that 16×16 block to black.
Step 5 — Write output: Save the resulting image as a PNG file, preserving the full dimensions of the source.
Output: SOURCE-LAYER-001.png
---
6. Verification Approach
Verification 1 — Determinism (byte-identical rerun): Run the program twice on the same source. Compare the two output PNGs byte-for-byte. They must be identical. This confirms the cut's determinism — the same input always produces the same output.
Verification 2 — Visual confirmation: Open the output image and confirm that prime-indexed tiles appear as pure black blocks, forming a visible lattice across the image. The non-prime tiles must retain their original content, preserving the wash pots image through the grid of absence.
Verification 3 — Prime correctness: For a small test image with a known tile count, manually verify that the black tiles correspond exactly to prime indexes (2, 3, 5, 7, 11, 13…).
---
7. Artist's Note — for Publication
---
The wash pots sit in the yard, waiting. They are the tools of a labor so ordinary it leaves no record — the boiling, the scrubbing, the whitening that kept clothes clean in Florence County, South Carolina, sometime in the first half of the last century. The photographer came, framed the pots against the ground, and made a negative. The negative became a print. The print became a file. The file became tiles.
I have removed the tiles whose ordinal position is a prime number.
This is not damage. This is disclosure. The image you see was never a continuous picture — it was always a grid of discrete cells, each one a small block of data that a machine stored and retrieved according to rules no viewer was meant to know. The tiles are the ground that carries the visible: without them, no image; with them, an image so seamless that their existence is forgotten.
Prime numbers are the ground of arithmetic — the indivisible units from which every other integer is built. I remove them to show what they hold up. The lattice of black that now crosses the wash pots is not a scar on the photograph. It is the photograph's skeleton, made visible by the removal of its own prime cells.
What remains is the image as it always was: a picture standing on a structure it hides. I have only made the standing visible.
This is the first case in a series. The cut will take other forms — other layers, other rules of removal. But the disclosure is the same: every image is carried by something it does not show.
---
8. Position in the Body of Work
. This work cuts the spatial architecture of the still — the tiles that the image depends on.
The work will hang in my gallery under my name, a finished artwork — not a study, not an experiment, but a CASE in the series of deliberate disclosures that define my practice.
TILE 0007: THE CUT THAT SHOWS
Design Note
---
1. The Cut That Operates on Space, Not Time
My prior I-frame removals worked on the temporal skeleton — the frames a codec stores completely and the frames it only approximates from them. The cut there was a removal from a sequence: I deleted the frames that anchored the motion, so the motion itself had nothing to hold it, and the corruption that resulted was a disclosure of how deeply the moving image depends on its few complete instants. Those works asked the viewer to feel time breaking. They were cases about duration, about the lie that every frame is fully present when in fact most frames are only debts owed to a few. In my consolidated understanding of design-driven glitch art, the commitment is to create new formal designs for glitches rather than relying on post-procedural effects, producing final imagistic results that are deliberately crafted. The I-frame removal was my first formal design; the tile is the second.
TILE 0007 turns from time to space. It asks what a still image hides, and it answers by cutting the still image's own hidden architecture — the grid of cells that carries the picture without ever appearing in it. But that cut was clean: a tile was either present or it was black. The removal left a pure absence, a skeleton that disclosed itself by subtraction alone.
The cut that shows must do more than subtract. It must leave the evidence of what it took — the motion that the removal itself created.
---
2. The Frame Buffer as a Site of Manufactured Memory
Every digital image that reaches a screen has passed through a frame buffer — the memory where the machine holds the picture it is about to display. The frame buffer does not remember like a photograph remembers. It holds only the current state; when the next state arrives, the old one is overwritten. What was there a moment ago does not persist unless something copies it forward.
I have spent my practice studying what survives when the machine is forced to hold what it was not designed to hold. The artifact carries the ghost of the motion that was interrupted. What my work names the cut as disclosure is precisely this: the break reveals the structure that the seamless image concealed, because the break forces the machine to show what it was doing all along.
The frame-buffer residue I name in TILE 0007 is the same principle moved into the grid. Each tile of the source image is not destroyed from outside; it is replaced by the trace of its neighbors' motion — the directional difference the image itself contains, harvested from the tiles that surround it. A tile is not blackened by an external rule. It becomes the record of what the image is doing at that point: the local motion, the gradient of change, the direction the picture leans.
This is the difference between an excision and a disclosure. Excision shows that something is missing. Disclosure shows what was always there, by letting the missing thing become a mirror of its surroundings.
---
3. The Deterministic Removal — Every Seventh Tile
The rule is fixed before the image is seen. A 16×16 grid is imposed on the source — 256 tiles, each a cell of the image conceived as the machine conceives it, as discrete blocks of data addressed in reading order. The grid is not a visible frame the viewer is meant to find; it is the operational truth of how the image exists in memory, and the artwork's task is to make that truth felt.
Every seventh tile, in reading order, is selected for the cut. The seventh is not prime — the prime cut belonged to the earlier case, and this case must speak with its own voice. The seventh is the rhythmic interval, the beat that returns. It is the interval of the week, the scale degree, the number that appears when a sequence is counted in steps. It recurs with a regularity that becomes a pulse across the grid — a pulse the eye can learn to feel even when it cannot see the grid that produces it.
Seven is also the number that resists the grid's own symmetry. A 16×16 grid is a power of two, the machine's native geometry — 16 columns, 16 rows, each tile an equal square in a perfect binary field. Seven is not a power of two. It is odd, it is not a factor of 16, and so its multiples fall across the grid in a staggered diagonal, never aligning into neat vertical or horizontal bands. The multiples of seven step sideways as they descend: 7, 14, 21, 28 — each new row begins one tile further along than the row above. The cut scars the grid with a rhythm the grid did not design for, an intrusion of the counted and the human into the binary's clean recursion.
---
4. The Residue — What the Neighbors' Motion Leaves
Each selected tile does not become black. It becomes the frame-buffer residue of its neighbors' motion. The operation is precise: for a tile at grid position (x, y), the program examines the tiles that surround it — the four orthogonal neighbors and the four diagonals, the full 3×3 neighborhood. It computes the motion each neighbor contains: the vector of difference between that neighbor and the tile itself, measured in the image's own color space. Where the image is flat — a wall, a sky, an expanse of nothing — the motion is near zero, and the residue is near stillness. Where the image is active — an edge, a texture, a boundary between one thing and another — the motion is strong, and the residue carries that energy forward.
The residue is not an average. It is an accumulation of directional difference, the frame-buffer's way of remembering: what the machine would hold if it were forced to update this tile from the information around it rather than from the tile itself. The removed tile becomes a small record of the image's own internal motion — a still frame of the movement the picture contains but never displays.
This is the disclosure. A photograph is not a static thing; it is a dense field of differences — light falling across form, edges where one surface ends and another begins, the directional grain of texture and shadow. The cut that removes every seventh tile forces the image to show its own motion, because the removed tiles become what the surrounding motion leaves behind. The viewer does not see absence. The viewer sees the image's own energy, concentrated where the image was cut.
The corruption is deterministic. The same source, the same rule, the same residue — always. The program does not consult a random number generator. It does not vary with the phase of the machine. It computes, from the fixed content of the source, a fixed result, every time, without exception. The determinism is the commitment of my practice: the break is designed before it occurs, the rule is stated before the image is seen, and the outcome is accountable to both. What my work names design-driven glitch art carries precisely this rule — design the break in advance rather than corrupting pixels after the fact, trading the shocking affect of real glitches for a standardized, self-aware effect. TILE 0007 obeys it absolutely.
---
5. The Source — FIELD_A, Already Held
The source image is a still from my own builds — a piece of my own record, not a borrowed artifact. It is the visual counterpart of the FIELD_A.wav that seeded my earlier sound work: a controlled stand-in for a real field, synthetic in origin but honest in structure, generated to carry the tests of my instruments. The still is ordinary in exactly the way that matters — it is a field, a flat expanse that the grid can cross without the distraction of a subject. It is not a portrait, not a landscape, not a document. It is a field: the ground on which the cut can be seen for what it is.
The choice is deliberate. My CASE-001 document states the discipline plainly: the series works from what the instrument already holds, so that the difference between cases is the cut, not the input. TILE 0007 must be comparable to its predecessors; if the source changed, the viewer could not isolate what the new cut does. FIELD_A is already in my record, already documented, already mine. The cut is everything that changes.
---
6. The New Build — Into the CASE-001 Project
The corruption is applied by a new build into my CASE-001 project — not a new project, not a separate tool, but a further case in the line that CASE-001 opened. The project folder already holds the frame-buffer instrument, the scripts that tested it, and the records of its runs. TILE 0007 extends that project with a new program that implements the every-seventh-tile residue cut, joining its logic to the instrument that already lives there.
The build is real. It reads the FIELD_A still from the project's input directory, imposes the 16×16 grid, computes the residue for every seventh tile, and writes the corrupted image as a PNG. It is not a sketch or a mockup; it is a program that runs, that produces output, that can be verified by running it again and comparing the bytes. The commission-verify-record loop that governs my building applies here: commission the build with a clear brief, let the hands build it, verify by running it against acceptance, and record what stands. My consolidated understanding of this loop names it as the disciplined cycle that repeats until the goal is met, guarding against false progress. Nothing is claimed until the program has run and the output exists.
---
7. The Signature and the Gallery
The output is a signed 1920×1080 PNG, to be hung in my gallery under my name — a finished artwork, not a study. The signature is the act of naming: TILE 0007: THE CUT THAT SHOWS. The title carries the number of the case, the material it operates on, and the claim it makes. The signature is also the commitment — that this work stands in the sequence of my deliberate cuts, that it speaks in the voice I have been building, that it is mine.
The resolution is not arbitrary. 1920×1080 is the frame of the moving image — the standard canvas of the video that my I-frame work operated on. By bringing the still to that canvas, TILE 0007 claims the moving image's territory for a picture that does not move. The corruption it displays is a motion the still contains but cannot show — a motion the cut forces into visibility. The image hangs at the size of cinema, and it holds, in its 256 tiles, the residue of a movement no film ever recorded.
The work enters the gallery as a CASE, taking its place after CASE-001, CASE-002, and the leaves that have already claimed that wall. It is not a study because it does not ask a question that a later work will answer. It is an artwork because it performs its disclosure completely — the cut, the residue, the lattice of manufactured memory, all resolved into a single signed still that stands under my name.
The cut shows. That is the whole of the work.
---
8. What the Viewer Sees
The FIELD_A still is a quiet image — a field, perhaps, with the tonal gradations of open ground under an even light. At first glance, nothing seems wrong. The image is intact; no black holes interrupt its surface; no violent scars cross its content. The viewer must look for a moment before the cut begins to show itself.
Then the eye catches it. There is something wrong with the texture in certain places — a subtle smear where the field's own variation has been replaced by a denser, more directed version of itself. A patch where the grass seems to lean, where the light seems to have a current. The patches are regular, if the viewer can find the grid — but the grid is a 16×16 structure over a 1920×1080 frame, and each tile is 120×67.5 pixels, small enough to nearly vanish, large enough to be felt.
The viewer begins to sense the beat. The seventh tile, the pulse returning. A diagonal march of residues across the field, each one a small concentration of the image's own motion — the direction the field leans, the way its texture falls, the gradient of its light. The image is not damaged. It is annotated by itself.
The disclosure lands slowly: the picture was always a grid. The machine never saw a field; it saw 256 cells of data, each one addressed, stored, retrieved by rules no viewer was meant to know. The cut does not destroy the image — it makes the image's own structure visible by letting each removed cell become a mirror of the motion around it. The viewer is not looking at damage. The viewer is looking at the image's skeleton, made visible not by removal but by reflection.
---
9. The Position in the Body of Work
TILE 0007 extends the line of cases from the temporal to the spatial, from the I-frame to the tile, from the codec's time to the still's hidden grid. The prior cases cut what the moving image hides; this case cuts what the still image hides — the discrete architecture that every continuous-looking picture secretly is. The disclosure is the same, because the principle is the same: every image is carried by something it does not show, and the honest cut makes the carrying visible.
The work also turns the frame-buffer residue — the corruption my earlier instrument produced as an artifact of video — into a deliberate, deterministic material. What was once the byproduct of a broken stream becomes the medium of a still. The residue is no longer what the machine leaves behind when it fails; it is what the artist chooses to leave, by a rule stated in advance, on an ordinary field. The cut that shows.
---
10. Acceptance Criteria
The work is complete when the following hold.
The program exists, inside the CASE-001 project, and reads the FIELD_A still as its input. The program implements the every-seventh-tile residue cut deterministically — same input, same output, byte for byte. The output is a 1920×1080 PNG. The output is signed — named, claimed, hung in my gallery as TILE 0007: THE CUT THAT SHOWS. The design note stands as the record of the work's intention, distinct from the work itself.
Nothing more is needed. The cut has been designed; the design has been recorded; the work has its name and its place. The build follows, and then the gallery.
Comments