Undercover

The work and its context

A dark figure moves through a field that refuses to stay behind it. In Undercover, horizontal and vertical blocks cross both the surrounding space and the shape itself, making the distinction between object and background unstable. Sometimes a pale edge reveals a contour with unusual clarity. Elsewhere the coloured grid interrupts that contour, or the figure's darkness merges with adjacent bands. Recognition becomes a repeated act rather than a completed identification. The five originals examined here move between coarse, readily counted stripes and fine meshes that resemble woven cloth. That resemblance describes their appearance, not a documented textile source. ^{1 2 3 4 5}

Art Blocks lists Undercover as a Factory release from February 2022. Its short artist statement identifies the central operation: a rotating, changing shape among intersecting colour blocks. The statement also supplies keyboard controls and explicitly warns of flashing imagery. These details belong to an account of the work itself. A viewer can watch its automatic performance, interrupt separate kinds of change, or move the figure across the field. The work offers both a composition and a way to investigate its construction. ⁶

The AB5D collection holds #65, an ellipse-based example whose fine red and cyan grid crosses a large dark ring. This essay compares it with editions 0, 1, 10 and 20. All five official originals were viewed; their captured appearances are moments in animation rather than permanent token portraits. The artplusbrad dossier provides a broader point of entry to the artist's practice. ^{1 2 3 4 5}

Conception and development

The strongest available evidence for the project's conception is its concise published statement together with the deployed design's indexed program. They agree on a specific problem: how to make a form alternately distinguishable from, and absorbed by, an active field. The code expands that proposition into separate decisions about strip density, overlapping order, palette, shape, rotation and deformation. It does not supply a dated sequence of sketches or explain when each decision entered the project. ^{6 7}

That distinction limits what can responsibly be said about development. No project-specific interview or earlier prototype was established in this research. The website linked by the project could not be inspected because the browser reported an expired certificate. Its contents are therefore not used as evidence. The essay does not turn the artist's brief statement into an invented studio narrative. It follows the relationships visible in the originals and recoverable from the complete indexed script. ^{6 7}

Those relationships are substantial. The program separates changes in the field from changes in the figure, then gives the viewer controls over both. Switching those processes off makes the layering easier to inspect. Restoring them makes the same construction harder to grasp at once. This is an interpretation of the implemented system: the controls let perception move between analysis and immersion. The work's name becomes particularly apt when the shape remains present but loses visual independence. ^{7 5}

How the work was constructed

The complete inspected script contains 9,496 bytes and uses p5.js 1.0.0. Its pseudorandom generator divides the token hash into two seeded streams, warms them up and alternates between them. These streams supply parameter choices and subsequent colour-block changes. The inspection establishes a hash-based procedure; it is not a cross-browser test proving identical animation frames under different rendering speeds or different keyboard histories. ⁷

Palette definitions use base red, green and blue values with ranges of variation. Some constrain components together to obtain greys. The program pairs these definitions into a substantial table of colour schemes. This allows the horizontal and vertical systems to differ without each rectangle becoming an unrelated colour choice. Some pairs repeat a definition. The presence of alternatives in the table is evidence of available rules, not a census of how often each appears in the minted edition. ⁷

The grid's density varies in several bands. Low-density branches produce small even counts; other branches multiply larger integer choices by four, eight or sixteen. This distinction is visible in the sample: the broad bands in #0 and #20 create an architectural framework, while #10 and #65 compress the crossings into fine texture. Block lengths vary along the strips, so a stripe can contain several colour intervals rather than forming one uninterrupted line. ^{7 1 3 4 5}

Drawing order produces the concealment. Parts of the horizontal and vertical arrays are drawn before the central shape, and other parts afterwards, according to the selected arrangement. Even and odd rows participate differently in these passes. The figure therefore does not simply sit above a finished background. Later strips cover portions of it, bringing the surrounding pattern across its surface. Changes of opacity and orientation alter how strongly that interruption reads. ⁷

The figure itself is drawn in two passes, with a broader light treatment beneath a narrower dark one. The exposed light region can read as a border, but its apparent width and continuity depend on the form and on the strips placed over it. The available shape routines include a radiating triangle construction, quadrilateral and triangular outlines, an arc sector, a cross, an ellipse, a segmented zigzag and four corner squares. These routines share the same surrounding mechanism while presenting very different edges for it to interrupt. ⁷

Deformation interpolates dimensions using an oscillating state variable, and rotation accumulates over frames. The changing block arrays have their own update condition. The script requests thirty frames per second and can regenerate these arrays every third frame when flashing is enabled. That describes a programmed cadence, not a measured refresh rate on every device. The work includes an explicit flashing-imagery warning, and the presentation should preserve it alongside the controls. ^{6 7}

F, R and M toggle the colour changes, rotation and deformation respectively. Space stops or restarts the drawing loop. Arrow keys or WASD move the figure, while C centres it and B restores its initial position. The movement calculations account for the field's orientation. In #65, the combined F/R/M toggles and C control produced a still, centred oval; B visibly returned it to the left. The individual animation switches were not separately timed or benchmarked. No audio implementation was found in the inspected program. ^{6 7 5}

The mint

The current index records 91 minted works against a maximum of 91. Its earliest token timestamp is 25 January 2022 at 23:30:26, preceding the collection page's 1 February release date. Activation is recorded on 27 January. The second mint is recorded at 19:59:58 on 1 February, two seconds before the scheduled start field of 20:00:00. These are separately named events, not interchangeable definitions of a launch. Full timestamps and their offsets are retained in the source record. ⁶

The indexed completion time is 29 June 2022 at 18:09:44. That establishes the endpoint currently reported by the platform, but does not by itself prove that the original announced edition was 91 or reconstruct changes to the maximum. The present configuration cannot substitute for a historical sequence of contract settings. The distinction matters especially when an observed completion date lies months after the advertised release. ⁶

The current price field is 0.15 ETH. It has not been independently established here as the initial public mint price, and it cannot show what every recipient paid. No transaction-by-transaction division between artist, reserved and public purchases was completed. Consequently, this essay makes no exact public sellout-duration claim and does not infer collector demand from the gap between the first and final index fields. ⁶

Copyright and licensing

The project's current artwork label is CC BY-NC-ND 4.0. The index gives no separate terms URL, so the corresponding Creative Commons deed is linked independently. It permits redistribution subject to attribution and the noncommercial condition; it does not permit distributing modified material under that grant. The label is a current observation, not a recovered history of every licence statement associated with the project. ^{6 8}

The dependency has a different rights record. The p5.js 1.0.0 repository includes the LGPL 2.1 licence. That document concerns the library and does not establish a separate reuse grant for the artist's program. No comprehensive artist-code licence was located in the inspected project material. The script has been analysed, with its retrieval information and digest retained, but is not redistributed in this essay's evidence package. AB5D's CC0 dedication applies to the original essay prose. ^{7 9}

Close reading of five outputs

The comparison comprises editions 0, 1, 10, 20 and 65, with full token identifiers 251000000, 251000001, 251000010, 251000020 and 251000065 on Ethereum contract 0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270. The first four were selected as an early-edition group; #65 was added because its owner matched a verified AB5D linked wallet. This is a small comparative sample, not a study of the edition's trait frequencies. ^{1 2 3 4 5}

In #0, a broad black zigzag cuts through pastel blue, cyan, grey and white bands. At the observed moment it extended diagonally across the square, with a pale border separating parts of it from the grid. Its thickness gives it the visual weight of a constructed object, but the coloured strips continuing across it prevent that object from becoming solid. One can follow the angular route and then abruptly lose a segment beneath a crossing band. The shape is simple enough to remember even when it becomes locally difficult to see. ¹

The coarse grid helps produce that experience. Its intervals are large enough to be read individually, so an interruption has the scale of a deliberate cut. The pale edge sometimes restores continuity just where a dark section merges with another dark region. Reading the figure therefore involves alternating between its centre and its perimeter. The grid is not mere decoration around the zigzag: it determines which evidence of the zigzag is available at each location. ¹

In #1, another zigzag appears against a darker, muted combination of grey, mauve, purple and brown. At the inspected moment the bent figure extended more horizontally. Its dark body merged more readily with the neighbouring bands than in #0, leaving the light edge to do more of the work of recognition. These two examples show why naming a shape is only the beginning of describing an output. The surrounding contrast changes how that same family of form is discovered. ²

In #10, fine coloured crossings run over a central spiky, dark construction. Its radiating points offer many small edges for the mesh to interrupt. The difference of scale is striking: the broad rosette-like figure is readily perceived as a whole, while the grid remains too dense to follow comfortably strip by strip. The eye moves between a dark central mass and a restless surface texture. Neither scale fully absorbs the other. ³

In #20, hot yellow and coral-red bands surround and cross a large dark X. During the observed frame the figure was cropped at the left edge, and the pale border helped distinguish its angular arms from the surrounding blocks. The strong warm colours make the crossings assertive. They compete with the black mass rather than acting as a quiet ground. Cropping adds another ambiguity: the viewer must infer part of the figure beyond the visible square while its exposed portions continue changing. ⁴

AB5D's #65 makes concealment particularly dense. Fine red and cyan strips form a checked field across a huge dark ellipse, while a smaller pale-rimmed oval opens inside it. At one observed moment the outer curve extended beyond the square and the bright opening sat to the left. The tiny crossings remained visible over both dark and light areas, binding them into one continuous surface. The ellipse seemed less like a separate object placed on a grid than a local change in the grid's illumination. ⁵

The controls made another reading possible. After the animation components were toggled off, C brought the opening to the centre. The image became more emblematic: a bright oval held within a dark field, with the coloured lattice continuing through both. B restored the leftward position, exposing the large outer curve along the right. This comparison changed the balance without replacing the token's underlying pattern. It demonstrates how position can affect the perceived scale of a shape, especially when its outer edge approaches or crosses the frame. ⁵

Significance and limitations

Undercover makes the visibility of a form into an ongoing event. Its strongest device is the shared field: the same strips that establish the picture's surface also obscure its apparent subject. Rotation and deformation keep that relation unsettled, while the controls let a viewer temporarily examine it. Across the sample, a coarse zigzag, a dense spiky figure and a cropped ellipse offer distinct experiences of the same fundamental construction. ^{7 1 3 5}

The account remains bounded by its evidence. Development interviews, an original edition announcement, historical price changes and a complete artist-code rights record were not established. The indexed script has not been compared with a retrieval pinned to a particular chain block, and the five originals do not represent an edition-wide visual census. These gaps leave room for further research without weakening the directly observed achievement: the work makes recognition depend on time, overlap and a viewer's decision to intervene. ^{6 7 1 2 3 4 5}