Pixel Glass

The work and its context

Pixel Glass makes a square behave like a surface catching light. Pale diagonal bands cross blue or pink fields, their edges appearing soft at a distance and stepped under closer attention. There is no pictured room, window frame or object behind the glass. The resemblance comes almost entirely from changes of brightness across a restricted grid. In the four outputs examined here, the same diagonal direction sustains markedly different balances between luminous opening and coloured surface.¹²³⁴

Kai released the series through Art Blocks Playground on 24 February 2021. The platform identifies a fixed edition of 256 works. Its brief description emphasises pixellated glass rather than offering a narrative or a detailed technical explanation. The work follows 27-Bit Digital, but replaces that project's conspicuous segmented figures with a quieter field. The kai artist dossier provides the broader context for the practice.⁵

This simplicity is consequential. With little to name in the picture, attention moves toward the width of a light band, the weight of a blue interval, or the point at which a regular cell becomes visible. The image can initially seem almost generic, like a decorative pane. Comparison reveals that its identity depends on the arrangement of transitions, not just the selection of a tint.¹²³⁴

Conception and development

The artist's launch post gives Pixel Glass two related roles. It is an independent artwork and collectible, while also providing a route to a physical glass print of 27-Bit Digital. The historical offer asked a participant to own a 27-Bit work and transfer a Pixel Glass token to the artist, with shipping arrangements handled separately. The announcement documents an intended exchange; it does not establish how many prints were supplied, whether a transferred token was destroyed, or whether the offer remains available.⁶

An earlier post explains the material interest behind that proposal. Kai connects physical and digital objects to previous work with embedded chips and printed electronics. Test glass prints for participants in the earlier project's development exposed the cost of shipping. A proposed universal gift became a more limited offer for early supporters. A later update to that post closes the promotion at the end of March 2021 and points readers toward the Pixel Glass exchange. That update belongs to the evolving page, not necessarily to its original February publication.⁷

The documented connection gives the title an additional resonance. A digital picture suggesting glass participated in an exchange involving an actual glass object carrying a different image. The essay's interpretation is that the series joins material suggestion to collecting behaviour. That relationship should not swallow the picture itself: the bands and colour deserve scrutiny independently of a historical benefit attached to holding the token.⁶¹³

How the work was constructed

The complete indexed script is short enough to follow from loading the page to drawing the last cell. It uses the browser's two-dimensional canvas directly. On load, it sizes the canvas to the page body, then takes the smaller dimension as the side of its drawing square. Dividing that side by sixty-four establishes the cell size. There is no external drawing-library call in the inspected script, and no need to imagine a three-dimensional glass object behind its appearance.⁸

The token hash determines the palette and progression. A final hexadecimal digit of zero selects the branch named Infused. That branch permits a much wider hue range and begins on a warm, pale ground. The ordinary branch uses a narrower range around blue and a cool, pale ground. Two preceding hash bytes help determine hue and starting lightness. These are specific relationships in the code, not a statistical survey of how often the finished edition exhibits each appearance.⁸

A small changing state controls the brightness of successive bands. The program reads hexadecimal characters, converts each to a number and subtracts eight. Its sign helps choose whether a running value rises or falls; its magnitude helps determine how long that direction persists. Additional bounds reverse direction when the value travels too far. Lightness is calculated from this value and constrained between limits, while opacity rises in part of the range. The resulting colour depends both on the foreground calculation and the ground beneath it.⁸

Crucially, the program changes that state once per diagonal, not once per cell. It first draws diagonals spreading outward from one corner, then completes the diagonals across the remaining half of the square. All cells on a given diagonal share the same calculated colour. The two passes cover 4,096 cells with 127 diagonal updates. This explains why the eye reads sweeping parallel movement even though the drawing primitive is a small axis-aligned square.⁸

The apparent softness is not produced by an inspected blur filter. There is no calculation of reflected rays, transmitted light, or an object distorted through a lens. Incremental colour changes across stepped strips create the suggestion of illumination. Close viewing can recover the cells that a broad view blends into a continuous surface. The technical economy is central to the effect: a modest sequence of colour decisions stands in for a much more elaborate material encounter.⁸¹⁴

A separate metadata function supplies the Infused and Tint names. Those labels describe selected parameters, but do not fully describe the arrangement of bands. Two outputs can share a tint while distributing brightness differently. The script also contains no animation loop or pointer interaction. Pixel Glass is a still work in its normal rendering, although its appearance can change perceptually as viewing size and distance change. Preserving the hash, script and canvas behaviour remains useful even when a single image seems sufficient to represent it.⁸¹²

The mint

The collection page dates the release to 24 February 2021. The index records the first mint at 05:09:24 that day, activation at 05:33:22, and the second mint and scheduled-start field at 13:40:58. Completion is recorded at 14:02:02. These fields describe distinct events. Their exact offset-bearing timestamps are retained in the evidence record, while the header uses the first mint's calendar date.⁵

The current project record reports 256 invocations against a maximum of 256. It also contains a price setting of 0.25 ETH. Neither that setting nor a subtraction of two timestamps establishes the entire public sale history. A reconstruction would need to distinguish artist or administrative mints from public purchases and locate the configuration applicable to each transaction. This essay therefore does not convert the recorded interval into an exact public sellout claim or the current price into a verified opening price.⁵

The launch post supplies a separate contemporary anchor: the artist announced the new project on the same date and described the print exchange. It helps establish how the release was presented, without replacing the contract and transaction evidence needed for a fuller mint history.⁶

Copyright and licensing

The indexed artwork licence field says NIFTY License and supplies no operative terms link. That label is recorded as observed. The specific text, version and history of the grant have not been established from the inspected project material. It would therefore be inaccurate to turn the label into a claim that the artwork is public domain or that unrestricted commercial reuse is permitted.⁵

The artist's script is a separate rights question. No independent reuse grant was identified in the inspected code. Its use of native canvas operations does not make the artist's program unprotected, and the absence of an external drawing library does not answer the artwork question. The historical print offer also concerns a particular exchange, rather than a general licence to reproduce the series. AB5D's CC0 dedication applies to this original essay prose; the artwork and artist code retain their separate rights status.⁸⁶

Close reading of four outputs

The comparison uses editions 0, 1, 10 and 20, selected as a small early-edition group before inspection. Their full token identifiers are 24000000, 24000001, 24000010 and 24000020 on Ethereum contract 0xa7d8d9ef8d8ce8992df33d8b8cf4aebabd5bd270. All four official previews were inspected as complete still images. This selection supports comparison, not a claim to represent the entire edition's distribution.¹²³⁴

In #0, a pale cyan field is crossed by several luminous diagonals. The differences are restrained enough that the picture first reads as one cool surface. Looking longer, the brighter lanes begin to separate and establish a rhythm across the square. The cell structure is present without taking command. Its interest lies in the threshold between recognising a grid and losing that grid in the impression of light.¹

In #1, brightness gathers much more strongly toward the lower-right portion. The blue field at the top and left feels more continuous, while the pale opening expands into a broad near-white area. Both #0 and #1 carry the Cobalt label, yet they do not distribute visual weight in the same way. The comparison demonstrates why a feature name cannot substitute for looking at the actual output: the transition between colours matters as much as their family.²¹

In #10, salmon pink surrounds cream-coloured bands. A broad pale diagonal travels from the lower-left region toward the upper-right, accompanied by thinner transitions. The warm field makes the result feel less like an icy pane than its blue neighbours. Its Infused and Lipstick metadata correspond to a different palette branch, but the familiar stepped direction keeps it within the same procedural family. The change of colour alters the emotional temperature without replacing the construction.³⁸

In #20, cooler blue and lavender tones produce more insistent alternating ribbons. The contrast makes the small cell structure easier to notice than in #0. A pale band can be followed as a continuous sweep, then broken mentally into the square marks along its edges. This oscillation between strip and cell prevents the work from settling into a completely smooth gradient. Its surface is both atmospheric and visibly assembled.⁴

Significance and limitations

Pixel Glass turns a simple grid into a study of how little information is needed to suggest a material surface. Its diagonals are not descriptions of a particular pane. They are sufficient cues for the viewer to supply a sensation of reflected light. The strongest comparison is between outputs that initially appear almost interchangeable: differences in the spread of brightness gradually reorganise the whole square.⁸¹²

The documented print exchange places that restrained visual experiment beside an early attempt to connect a generative token to a physical object. The work's historical interest lies partly in that connection, while its visual interest remains available in the bands themselves. Neither a utility label nor a rarity label accounts for the experience of looking across the series.⁶⁷³⁴

Unresolved research questions:

• Recover the original price and edition announcement, including any configuration changes.
• Distinguish public purchases from artist or administrative mints.
• Locate the operative NIFTY License text and any separate artist-code grant.
• Establish the history and fulfillment of the glass-print exchange without assuming current availability.
• Compare the indexed script with a pinned chain retrieval and broaden inspection beyond the four early outputs.