Hold a magazine photograph under a strong magnifier and it quietly falls apart. The seamless gradient of a face or a sky resolves into thousands of tiny colored dots - cyan, magenta, yellow, and black - each one crisp and separate, arranged in disciplined grids that tilt at slightly different angles depending on the ink. Back away to reading distance and the dots vanish back into the image they were pretending to be. This is halftone printing, and it is one of the most successful illusions in everyday technology: nearly every photograph you have ever seen on paper, from newspapers to postage stamps to cereal boxes, is a dot construction. Your screen does something related with its tiny red, green, and blue subpixels, which a droplet of water will happily magnify into visible stripes, but the two systems arrive at the trick from opposite directions. The screen builds color from lights you cannot individually resolve; the press builds tone from ink spots you were never meant to see.
Four inks that cannot whisper
The problem the halftone solves is brutally physical. A printing press lays down ink, and for any single ink the press essentially has two comfortable states: paper with the ink on it and paper with the ink off. Darkening a region means more ink, but a photograph is made of thousands of distinct brightness levels, and presses cannot meter ink in thousands of subtly different doses at photographic resolution. The workaround is to stop trying to vary the ink and vary the area instead. Each tiny region of the image becomes one dot decision: a big dot for a dark area, a pinprick for a light one, no dot at all for bare paper. From a distance the average of dark dots and white paper across any patch of the image lands somewhere on the gray scale, and the eye, doing the averaging without being asked, reconstructs the photograph. Engineers have refined the dot-placement mathematics for over a century - first with photoengraved screens, later with digital algorithms that compare the image against mathematical threshold patterns to decide where dots should fall.
Your eye closes the gap
The entire illusion depends on one property of human vision: its limited resolving power. At normal reading distance, dots smaller than roughly a tenth of a millimeter blur together, because your eye can no longer separate objects that fall within the span of a single receptor's receptive field on the retina. Print buyers exploit this deliberately. A newsstand magazine printed for close inspection uses a fine screen of perhaps 150 dots per linear inch; a billboard viewed from thirty meters gets away with dots you could count individually from the shoulder of the road, because at that distance even chunky dots sit below your eye's resolution limit. The viewing distance and the dot size are chosen together, which is why the same photograph can look silky in one print and pebbled in another without the image file changing at all.
Why the grids run at angles
One more piece of engineering hides inside the dot grids, and you can find it with the same magnifier. The cyan grid, the magenta grid, the yellow grid, and the black grid are each rotated to a different angle - commonly close to forty-five degrees for black and gentler tilts for the colors. The rotation is not decoration. When two regular dot grids overlap at nearly the same angle, their rows beat against each other like misaligned window screens, producing large interfering patterns that crawl across the image. Printers call it moiré, and it is the same family of beat pattern you create whenever you photograph a screen with another screen and wavy bands appear. Setting each ink's grid at its own angle spreads the interference out into a small, uniform rosette texture instead of great ugly bands, which is exactly what you see under the magnifier: tiny flower-like clusters repeating everywhere the inks overlap.
Where the dots give up
Knowing the trick also tells you where it fails, and once you notice these failures you will see them everywhere. Very pale tones are the first casualty: a tint so light that its dots become sparse and widely spaced starts to look grainy, because there are too few dots left for your eye to average into a smooth wash. Deep, saturated tones strain the system from the other end, since the total ink a press can pile onto paper has limits, and the most luminous colors a screen can emit - the neon greens and hot pinks - simply have no dot recipe that reproduces them on white paper with reflective ink. Smooth wall-paper gradients are the everyday reminder: when a printer or a low-quality display quantizes a gentle sweep of color into too few levels, faint stair-steps appear in exactly the places the dots and codes ran out of intermediate values. None of this means the print is broken; it means the illusion has reached the edge of the resolution, ink, and arithmetic that sustain it. The rest of the time, the dots hold their silence - and a photograph made of four colors and paper keeps passing as a window.