A photographer exports a set of images, uploads them to a gallery, and opens the page on a laptop. The reds look dull, the skies flat, the skin tones slightly grey. On the editing screen they looked rich. Nothing went wrong with the files themselves. The colour profile did.
Every digital image describes colour as numbers, and a colour profile tells software what those numbers mean. The same value for “full red” means a more saturated red in a wide colour space like Adobe RGB or Display P3 than in sRGB. If the image is saved in one space and shown as if it were in another, the colours shift. Usually they go dull, because a wide-gamut image read as sRGB loses saturation.
Why sRGB is the safe default
sRGB was designed in the 1990s as a standard space for monitors and the web, and it remains the assumption almost everywhere. Browsers treat untagged images as sRGB. Many content management systems, image plugins and social platforms strip embedded profiles when they generate resized copies, and the copies then get read as sRGB whatever they started as. Older and cheaper screens can barely show more than sRGB anyway.
So an image exported in sRGB looks close to the same on almost every device: the photographer’s calibrated monitor, a visitor’s office laptop, a five-year-old phone. An image exported in Adobe RGB might look excellent in one browser that reads its profile, wrong in a thumbnail that lost the profile, and wrong again on a page where a plugin re-encoded it. For a gallery, where dozens of images sit side by side, that inconsistency is worse than slightly narrower colour.
Where wider colour makes sense
Most recent phones, tablets and many laptops have wide-gamut screens close to Display P3, and modern browsers handle tagged P3 images correctly. For work where intense colour is the point, such as saturated flowers, neon at night or vivid product colours, a P3 version can look noticeably richer on those screens.
It only works if the whole chain keeps the profile. That means checking that your gallery software, image optimiser or CDN keeps embedded ICC profiles in every resized version, not just the original. Test it: upload a P3 image, then open each generated size and check its profile. If any step strips it, those versions will look duller than the sRGB file would have.
A practical compromise is to keep sRGB as the default for every image and use P3 only for a small set of images where the difference matters, on a gallery setup you’ve tested end to end.
A simple export routine
Edit in whatever wide space your editing software prefers. That keeps more colour information during editing. Then, at export:
Convert to sRGB rather than just assigning the profile. Converting remaps colours so they look as close as possible to the original; assigning just relabels the numbers and changes the look.
Embed the profile in the file. It costs a few kilobytes and helps the browsers that read it.
Check the result in more than one browser and on a phone before publishing a whole set. If an image clips badly in sRGB, with saturated areas turning into flat blocks, adjust that image’s saturation rather than switching the whole gallery to a wider space.
Last, don’t chase the look of one screen. A calibrated editing monitor shows what the file contains. Visitors will see it on screens that are too blue, too bright or set to a warm night mode. sRGB won’t fix their screens, but it keeps your gallery from adding a second problem on top.
For more on preparing images for galleries, see our guides to JPEG, WebP and AVIF and media gallery image sizes.