Color Study

The work and its context

Color Study makes a square into a field of relationships. Colours shift across a grid, sometimes in broad, countable steps and sometimes in intervals fine enough to approach a continuous blend. White rules can divide that field into vertical or horizontal strips. These lines do more than describe the underlying cells: they give the eye a direction to follow, making the same two-dimensional movement of colour read as a succession of bands. The four inspected outputs are orderly without being optically inert.¹²³⁴

Jeff Davis released the series through Art Blocks Playground in February 2021. The current collection page identifies 2,000 works and describes the interaction of four parent colours. The term is useful because most colours visible in an output are calculated relationships between starting values, rather than independently selected swatches. Its place in Davis's practice is an extension of his investigation of geometric order and colour; the Jeff Davis dossier gives the broader context.⁵⁶

The distinction between a colour chart and a picture remains productively unsettled. A chart promises information that can be extracted and used elsewhere. Here the arrangements themselves hold attention: a dull middle may connect brilliant opposing corners, while white separators compete with the colours they appear merely to organise. That reading arises from the selected works, not from a claim that the artist intended to illustrate a scientific measurement.¹²³⁴

Conception and development

Davis's February 2021 Art Blocks interview describes chance entering an existing practice of painting, printmaking and digital composition. Random-number tables in Excel initially guided decisions in Illustrator; Processing later combined the decisions and drawing. At the end of that interview he previews an unnamed Artist Playground project concerned with colour. Its timing is suggestive, but the passage alone does not identify Color Study by name or establish a precise development schedule.⁷

In a 2023 interview with Alex Estorick, Davis identifies Josef Albers as important to his understanding of art that demonstrates colour relationships through looking. He describes an approach based on integers, proportion and arithmetic rather than noise, and names Color Study as his largest project at that time. He also argues that unexpected perceptual effects can emerge when a limited set of colours is placed together. This is retrospective testimony about his method, not a launch announcement.⁸

The inspected program makes that connection unusually explicit: it contains an Albers branch that paints four nested squares. It does not follow that every output is a quotation of Albers. Most of the routine instead distributes mixtures across a rectangular grid. The reference exists as a particular coded possibility within a broader system of colour arrangement. No dated sequence of prototypes was recovered for this essay, so the order in which those possibilities were developed remains open.⁶

How the work was constructed

The retrieved artist script uses p5.js 1.0.0. It reads thirty-two pairs of hexadecimal characters from the token hash into numerical values and derives a separate seed from another span of the same hash. The first two values select horizontal and vertical divisions. Later values provide red, green and blue components for four parent colours and switches for alternative treatments. These findings come from reading the complete indexed script, not from independently executing a downloaded copy.⁵⁶

The division list runs from three to 120, with many repeated entries. Choices are therefore weighted by their repeated positions rather than spread evenly across the distinct division counts. Further rules prevent both axes from simply taking the sparsest arrangements together: when one receives three or four divisions, the other is selected from a denser portion of the list. This keeps extreme coarseness directional. The canvas is square, but its individual cells need not be.⁶

In the ordinary grid, each cell receives a weighted mixture of the four corner colours. Moving horizontally changes the balance between left and right; moving vertically changes the balance between top and bottom. The calculation is applied separately to red, green and blue. This is bilinear interpolation in practical terms: a colour's position determines how much each corner contributes. It is a numerical blend, not a simulation of mixing pigments. At a corner the corresponding parent dominates completely; intermediate cells combine several contributions.⁶

The program also edits its starting material. Very dark initial colours can be lifted. Other branches push colours toward complementary hues, assign one corner black, lighten the palette, or maximise saturation and brightness. A contrast check can reject the first group of parents and take a second group from later hash values, reapplying the selected treatments. That check explains why the system is more directed than four unrestricted random colours poured into a gradient.⁶

Order of operations matters. Some switches exclude others. A random-colour override replaces the normal mixture at each cell with seeded channel values. In the complementary branch, the fourth parent's assigned channels actually reuse the first parent's adjusted colour, even though another intermediate value is calculated. Describing only the variable names would miss that detail. These are inspected possibilities, not audited frequencies across the edition.⁶

Shapes and separators are handled after those choices. Equal row and column counts can allow circles, leaving white spaces between discs. Otherwise the routine paints rectangles with matching outlines. White horizontal and vertical rules can be added independently when circles are absent. Their width scales with the canvas. A dense set of rules can consequently become a substantial visible component, rather than an incidental border between colour samples.⁶

The Albers branch is triggered by particular four-character sequences within the hash. It paints its nested squares after the grid and lines, covering the earlier image with a different organisation of colour. The squares mix the first and fourth parents and descend as they shrink, producing unequal surrounding margins. This exception changes composition as well as palette. It is distinct from the four line-based outputs selected below.⁶

The drawing finishes with noLoop, and the retrieved script contains no animation or pointer interaction routine. Its output is static. Reproduction nevertheless depends on the token input, artist program and compatible p5 environment, with the canvas fitted to the smaller window dimension. The indexed retrieval used here was not independently compared against a pinned Ethereum block; the research record retains its digest and query location for that later check.⁶⁹

The mint

The collection page dates the release to 8 February 2021. The index observed on 17 September 2026 reports the first mint at 04:55:41 that day, activation at 05:08:20, and both the second mint and scheduled start at 20:55:53. Those fields describe different events. The early first token should not be silently treated as the opening of a public sale, nor should activation replace the observed mint time. Full timestamps and offsets are retained in the evidence record.⁵¹²

Completion is recorded on 21 February at 14:39:41. The current count and maximum are both 2,000. The intervening calendar period establishes that the indexed first and final events occurred on different days; it does not prove continuous availability, identify pauses or classify every recipient. The four inspected token records preserve individual mint transaction hashes, but they do not constitute an audit of the sale.⁵¹²³⁴

The current price setting is 0.05 ETH. A launch-era price announcement and a complete configuration history were not established in the inspected sources, so this value is not presented as the amount paid for every mint. Likewise, today's maximum and the current edition description do not independently establish the originally announced cap. No exact public sellout duration, gas-cost estimate or undocumented technical incident is asserted here.⁵

Copyright and licensing

The current artist-set artwork label is CC BY-NC 4.0, recorded here as creative-commons. The metadata provides no terms URL; the corresponding Creative Commons deed is linked separately. It permits sharing and adaptation subject to attribution, identification of changes and the noncommercial condition. This records the label observed during research, rather than a reconstructed history of every earlier licence statement.⁵¹⁰

The p5.js 1.0.0 repository contains the GNU Lesser General Public License version 2.1. That is a library licence, not evidence of a separate permission to redistribute Davis's artist script. No separate script grant was identified in the retrieved program or project metadata. The essay analyses that program without republishing it. AB5D's original prose is offered under CC0; artwork, artist software and third-party library rights remain separate.⁶⁹

Close reading of four outputs

Editions 0, 1, 10 and 20 were selected in advance as a small early-edition comparison. Their full token IDs are 16000000, 16000001, 16000010 and 16000020, on Ethereum contract 0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270. All four official static previews were visually inspected. Each has no indexed colour override or modification; this deliberately limited comparison therefore does not illustrate the full range of branches described above.¹²³⁴

In #0, blue occupies the upper left while green gathers at the upper right; the lower right turns coral and the lower left becomes a muted blue-green. Twenty-four white-lined columns make the field appear tall and narrow in its internal rhythm, even though the overall image is square. Six rows remain legible as broad steps within those columns. Near the middle, less insistent colours provide a passage between the brighter corners. The white lines continue at the same strength through each zone, refusing to let the quieter middle disappear into an undifferentiated haze.¹

This creates two competing ways to look. Following one column downwards reveals six changes of colour. Looking across the square gives a much slower impression of one family passing into another. The vertical lines privilege the first movement, but the overall blend encourages the second. The composition's apparent simplicity depends on holding those different viewing directions together rather than resolving them into one dominant figure.¹

In #1, green at the upper left meets violet at the upper right, while the lower edge moves from deep blue to luminous yellow. Forty columns and twenty-four rows divide the colour into much finer increments than #0. The bright lower-right region appears to shine outward through a comparatively subdued centre. Yet every colour sits on the same flat grid. The impression of illumination is produced by relationships of hue and value, not by a pictured lamp or a modelled source of light. The repeated white verticals both reinforce and interrupt that illusion.²

The horizontal lines of #10 change the dominant rhythm. Sixty rows create a fine white mesh across a movement from pink and purple above toward greens below. Twenty columns are less immediately conspicuous than the densely stacked separators. From a distance the colours seem to sit behind a screen; closer looking recovers discrete steps. Beside #1, the image demonstrates how rotating the emphasis of the separators changes the reading without requiring a different basic mixing rule.³

In #20, twelve horizontal bands make that structure much broader. The top crosses from warm red to brilliant green, while the bottom rests in blues. The white divisions provide intervals wide enough for the colour changes inside each strip to be followed separately. Compared with #10's dense interference, the result is calmer and more countable. The two works have twenty columns each, but the different row counts make their temporal experience of looking markedly different.⁴

Significance and limitations

Color Study gives the grid an active role in seeing colour. It controls how often change becomes visible and which direction the eye follows, while the corner mixtures tie distant parts of the square together. The four examples show that a shared construction can produce different apparent brightness, pace and depth without representing an external scene. Those effects are the essay's visual interpretation, grounded in identifiable works.⁶¹²³⁴

The artist's explanation of the later Progression explicitly identifies its tinted and saturated treatments with those used in Color Study. That is documented continuity in his vocabulary, not a claim that the projects share an identical algorithm. Color Study can be understood as one sustained formulation of a problem Davis continued to develop: how numerical decisions become perceptual relationships.¹¹

Unresolved research questions:

• Recover the original price and edition announcements, with any subsequent changes.
• Classify public, artist and administrative mints and identify any pauses from transaction history.
• Establish separate artist-code terms and the history of the artwork licence.
• Compare the indexed script with a pinned chain retrieval and a documented rendering environment.
• Extend visual comparison to circles, nested squares and palette overrides without treating this four-work sample as representative.