RGB to HSL Converter — Free, Instant

Grey colours return hue zero, which means nothing

When you convert a grey like rgb(128, 128, 128) to HSL, this tool returns hsl(0, 0%, 50%). The hue is mathematically undefined — the red, green and blue channels are equal, so there is no dominant colour and the formula would divide by zero. The code sidesteps this by leaving hue at 0, but that 0 reads as red everywhere else in the HSL system. If you programmatically read the hue back expecting it to mean something, you will be surprised. Always check saturation first: when S is 0, the H value carries no information at all.

Walking the conversion step by step

The algorithm starts by normalising RGB from 0–255 to 0–1 by dividing each channel by 255. Then it finds the maximum and minimum channel values. Lightness is the midpoint of these two extremes: L = (max + min) / 2. This is not perceptual — it ignores the middle channel entirely, which is the root cause of HSL's reputation for uneven-looking palettes.

Saturation uses a two-branch formula, and this is where HSL gets tricky. Below L = 0.5, saturation is S = (max − min) / (max + min). At or above L = 0.5, it switches to S = (max − min) / (2 − max − min). Why two formulas? Saturation is defined as chroma relative to the maximum chroma still possible at that lightness, and that ceiling collapses towards both ends of the scale — a colour at 95% lightness simply cannot be very far from white. The denominator has to shrink with the ceiling. Without the switch, colours near white or black would report artificially low saturation.

Hue depends on which channel is largest, and there are three cases. If green is max, H = (B − R) / d + 2. If blue is max, H = (R − G) / d + 4. If red is max, H = (G − B) / d, plus 6 when green is less than blue. That last conditional is easy to miss and it matters: with red as max and green below blue, the fraction is negative, which would put hue in the reds-through-magentas below zero. Adding 6 wraps it back around the top of the wheel instead. The result is then divided by 6 and multiplied by 360, giving a contiguous 0–360° range.

Precision: the round trip is lossier than you would expect

This tool rounds hue, saturation and lightness to whole numbers, and that rounding costs more than it sounds like. Sweeping 636,056 evenly spaced colours through RGB → HSL → RGB, 88.9% of them came back changed. Most shifted by 1 or 2 out of 255 in some channel, and the worst cases reached 5: rgb(3, 228, 234) becomes hsl(182, 97%, 46%) and returns as rgb(4, 223, 231). Plenty of colours are unaffected — rgb(123, 45, 67) converts to hsl(343, 46%, 33%) and comes back exactly — but you cannot tell in advance which yours is.

The cause is simply how few distinct values whole-number HSL has: 360 × 101 × 101 possible triples against 16.7 million RGB colours. For comparison, OKLCH reports four decimal places and drifts on well under a tenth of one percent of colours. So use HSL as a working format for reasoning about and adjusting colour, but keep the original RGB or HEX as your source of truth.