Put the word RED in bright red and the word BLUE in bright blue on the same black background, and many readers instantly see something odd: the two words do not lie on the same plane. The red floats a little forward; the blue settles a little back, like letters pressed into the screen's depth rather than painted on it. No glasses, no trickery, no actual depth — the screen is as flat as any surface you own. The effect is called chromostereopsis, and it earns its illusion honestly: it comes from the same optical machinery that would deliver real depth, fed a wavelength-based counterfeit instead. It is the quiet sibling of red-and-cyan glasses building depth from wavelength — that trick filters whole scenes, while this one needs nothing but two strong colors and your own two eyes.
A lens that treats colors unequally
Every lens, including the ones you were born with, refracts short wavelengths slightly more strongly than long ones. Blue light entering your eye focuses a hair closer to the cornea than red light does — the eye's own chromatic aberration, small but never zero. If your pupil were perfectly centered and perfectly symmetric, that would be the end of it: both eyes would suffer identical, symmetric blur, with no depth signal. But pupils typically sit a touch nasal of the eye's optical axis, and that small offset breaks the symmetry. Each eye images the red word and the blue word at minutely different positions on its retina, the differences are not quite the same in the left and right eye, and the visual system does with that discrepancy what it does with every such discrepancy it has ever encountered: it computes depth. The mechanism, quantified in modern studies as transverse chromatic aberration, converts a tiny wavelength-dependent geometry into a convincing near-and-far.
The display-era version
Screens made the phenomenon democratic. Experiments with flat displays show that the illusory depth grows as the color difference between regions grows — pure red against pure blue on a dark surround is close to a worst case, which is exactly the combination early computer graphics and slide designers loved. The same research found a detail that sounds impossible until you remember the geometry: shift the observer's pupils sideways, covering the nasal or temporal side of each pupil, and the depth can weaken, vanish, or flip, blue now standing in front of red. The effect you get depends on where the light passes through your pupil, which is why two people can look at the same slide and argue about which word is nearer.
If neither word floats for you, that is a normal outcome too — the direction and strength vary with pupil position and observer.
Not an eye problem — an eye property
Chromostereopsis has been in the vision-science literature for well over a century, reviewed and re-measured across generations of instruments, and it belongs to healthy eyes. It is not a symptom of anything, and noticing it strongly says nothing about your vision. What varies naturally among people is the sign and strength: most often red appears nearer, but some observers see the reverse, others see almost nothing, and the same person's experience can change with pupil size — dim room, wide pupils, stronger effect. Ophthalmology's older reviews dwell on exactly this person-to-person spread, which is one reason the phenomenon stayed a curiosity rather than becoming a tool.
The variability is a design constraint, not a footnote. A data visualization that encodes meaning in 'the red layer looks closer' will simply fail for the minority who see blue forward or nothing at all, and will change strength for everyone as room lighting changes their pupil size. So practical dark-theme guidance leans on the parts that do not vary: keep large red and blue fields from sharing a border at full saturation, match their perceived lightness rather than their drama, and reserve strong chromatic contrast for elements separated by space. The floating words become a diagnostic, of all things — a quick way to check which of your own eyes' habits a given color scheme is quietly leaning on.
Why designers quietly route around it
Anyone who has built a dark-theme interface or a nighttime presentation has met the practical face of chromostereopsis: saturated red text on saturated blue, or either on black, can shimmer, feel like it vibrates, or simply tire the eyes as they micro-adjust focus between wavelengths that refuse a shared focus. It is a cousin of the trouble behind white text seeming to glow on black backgrounds — again the eye's optics, not the design, supplying the weirdness. Experienced designers respond with physics-aware habits: avoid extreme red-blue pairings side by side at full saturation, keep lightness levels close, soften one of the colors, or add separation with space and less extreme hues. The choices are easy to rehearse by eye in Complementary Color, where strong opposing hues sit adjacent and you can feel which pairs rest easily and which push.
What the floating words prove
The strangest takeaway is the deepest one: color is not decoration laid over a 3D world — it participates in how depth itself is computed. A pair of wavelengths, unevenly bent, can counterfeit an entire dimension on a flat panel. When you next see a red logo riding above its blue background, you are watching your visual system run its usual geometry on slightly counterfeit inputs and produce, without complaint, a small, private, physically impossible space. That is the signature of an interpreter rather than a camera: given inconsistent evidence, it does not error out; it invents the most plausible world the numbers allow.