Every screen you have ever stared at is lying to you — not maliciously, but structurally. It can only show you a subset of the colors a human eye can perceive, and most people working in digital products never think about what sits outside that boundary. That gap has real consequences for designers, hardware engineers, and anyone building software that touches color-critical workflows.
The Triangle on the Diagram Nobody Reads
Color scientists represent visible light as a horseshoe-shaped region called the CIE 1931 chromaticity diagram. Every display standard — sRGB, DCI-P3, Rec. 2020 — is drawn as a triangle inside that horseshoe. The triangle is the gamut: the full range of colors a device can reproduce by mixing its three primaries.
The geometry is blunt. sRGB, still the default color space for most web content, covers roughly 35% of visible colors. DCI-P3, used in modern iPhone and MacBook displays, pushes that to around 53%. Even the ambitious Rec. 2020 standard, which almost no consumer display fully achieves in hardware, only claims about 75%.
The remaining slice — particularly the saturated cyans, blue-greens, and certain vivid yellows — simply cannot be encoded as a combination of any three real phosphors or LEDs arranged in a triangle. Physics, not engineering laziness, draws that ceiling.
Where the Missing Colors Actually Live
So where do you encounter colors that fall outside the sRGB triangle in the real world?
- Open sky near the horizon at solar noon. The saturated cyan of a clear tropical sky regularly exceeds sRGB and even P3 gamuts. Photographs of it are, by definition, compressed approximations.
- Bioluminescent organisms. Certain deep-sea creatures and fungi emit light at spectral purities that no display primary can match without clipping.
- Interference colors in thin films. The iridescent sheen on a soap bubble, a beetle's carapace, or an oil slick exploits wavelength-level interference — producing spectral purity that pigments and phosphors cannot replicate.
- Saturated laser light. A green laser pointer at 532 nm sits at a spectral locus point that falls well outside most display triangles. You can see it; you cannot photograph it faithfully.
- Some cut gemstones under directional light. The fire in a high-quality emerald or alexandrite under a single-point light source produces saturated hues that camera sensors clip before the image reaches your retina.
The common thread: these colors arise from narrow-band spectral emission or high-purity interference, not from broad reflectance. Displays work by mixing broad-band primaries, which is inherently a lossy compression of the spectral world.
Why Software Teams Should Care
If you build anything that processes, displays, or stores imagery, this is not just a curiosity.
Color pipeline correctness
Most image processing pipelines default to sRGB, which means wide-gamut source images get silently clipped during ingest. If your product handles photography, medical imaging, satellite data, or e-commerce product photos, that clipping may be destroying information your users paid to capture.
A minimal safeguard:
from PIL import Image, ImageCms
def ensure_wide_gamut_preserved(path: str) -> Image.Image:
img = Image.open(path)
icc = img.info.get("icc_profile")
if icc:
src_profile = ImageCms.ImageCmsProfile(BytesIO(icc))
dst_profile = ImageCms.createProfile("sRGB")
# Check before converting — don't silently flatten a P3 image
src_cs = ImageCms.getColorSpace(src_profile)
print(f"Source color space: {src_cs}")
return img
The point is not that you always need to preserve wide gamut — it is that you should make the decision deliberately rather than by default.
CSS and the color() function
The web platform now supports wide-gamut color natively. The CSS color() function lets you address Display P3 primaries directly:
.hero-cta {
background-color: color(display-p3 0.12 0.87 0.52);
}
Browsers on wide-gamut displays will render a green that sRGB literally cannot encode. Browsers on sRGB screens fall back gracefully. If your design system still hardcodes everything in #rrggbb hex, you are leaving visual fidelity on the table for the half of your users who own modern hardware.
HDR video and streaming products
If you are building a video platform, HDR10 and Dolby Vision content operates in the Rec. 2020 container specifically to preserve as much of the original scene's color as possible. Transcoding pipelines that strip tone-mapping metadata or force BT.709 output are not just technically incorrect — they actively degrade the product experience for users with HDR televisions.
The Deeper Engineering Lesson
The boundary of a display gamut is a good mental model for a class of problem that appears everywhere in software: lossy representation boundaries. Audio is sampled at a finite rate. Floating-point numbers cannot represent all rationals. UTF-8 historically struggled with certain scripts. In each case, the system works beautifully inside its design envelope and fails silently — or destructively — outside it.
The professional habit is to know where those envelopes are before a user hits them, not after.
Wide-gamut color is becoming a baseline expectation in consumer hardware. Apple, Samsung, and Google have shipped P3 displays across their product lines for years. The web platform supports it. The question is whether the software layer in between is keeping pace or quietly discarding data that users assume is being preserved.
Source: Where to Find the Colors Your Screen Can't Show You — https://moultano.wordpress.com/2026/06/19/where-to-find-the-colors-your-screen-cant-show-you/
Why this matters for your project: Whether you are building a media app, an e-commerce storefront, or a data visualization tool, color pipeline decisions made early in architecture have a way of becoming expensive to reverse. Auditing your image ingest, CSS color space assumptions, and export formats now — before your user base scales — is the kind of low-visibility, high-leverage work that separates products that age well from ones that quietly erode in quality.
