Grayscale is a style choice and a practical one
Black-and-white photography has stayed a deliberate creative choice for as long as color photography has existed, precisely because removing color forces attention onto composition, light, texture, and contrast instead of color relationships. But grayscale conversion isn't only an aesthetic decision — it's also genuinely practical: preparing an image for a context that only displays in black-and-white, simplifying a design mockup before color gets finalized, or creating a grayed-out, disabled-looking version of an icon or thumbnail, a pattern used constantly in app and web interfaces.
Whichever reason brings you to it, the actual quality of the conversion depends entirely on the math behind it, and not every tool gets that math right.
Why a flat average produces flatter results
The naive way to convert a color pixel to gray is to average its red, green, and blue values. This is quick to compute, but it produces results that don't match how human vision actually perceives brightness — our eyes are far more sensitive to green light than to red or blue at the same intensity, which means a flat average systematically under-represents how bright green tones should look and over-represents blue.
A proper weighted luminance calculation accounts for this directly, weighting green most heavily, then red, then blue least. The visible difference shows up most in images with strong color contrast that doesn't correspond to strong brightness contrast — a red apple next to green leaves, for instance, can end up nearly indistinguishable in tone with a naive average, while a properly weighted conversion preserves the actual brightness difference between them.
What you can't undo, and what's preserved
Grayscale conversion genuinely discards color information rather than just hiding it — there's no algorithm, however clever, that can accurately reconstruct the original colors from a black-and-white result, since the information simply isn't there anymore. This is worth remembering before overwriting your only copy of an image: keep the original color file if there's any chance you'll want it again.
What does survive the conversion intact is transparency — if your source image has an alpha channel (a PNG logo with a transparent background, for instance), only the color channels get converted; the transparency itself passes through unaffected.
Grayscale versus sepia and vintage filters — a common mix-up
People sometimes ask for "black and white" when what they actually picture is a warm, aged, sepia-toned look — these are different effects. True grayscale removes color entirely, leaving pure black, white, and gray tones with no color cast at all. Sepia and vintage filters apply a specific color tint on top of a desaturated (or partially desaturated) base image, which is a more stylized, deliberately warm-toned effect rather than a neutral grayscale conversion.