Lumens, Lux and Candela: Buying Lighting Without Being Misled
Three units describe light output, and manufacturers pick whichever flatters the product. Here is what each one actually measures.
Lighting is sold with a confusing mixture of units, and the confusion is not always accidental. A torch advertised at "100,000 lumens" is almost certainly quoting something other than lumens, or quoting an LED die rating that the finished product never delivers.
The three quantities
Candela (cd) measures luminous intensity — brightness in a particular direction. It is the SI base unit for light.
Lumen (lm) measures luminous flux — total light output in all directions. This is the figure that matters when replacing a bulb.
Lux (lx) measures illuminance — light arriving on a surface. One lux is one lumen per square metre.
Why lumens replaced watts
Watts measure power consumed, not light produced. With incandescent bulbs this was an acceptable proxy because efficiency barely varied. LEDs broke the proxy entirely.
Rough equivalents for replacing incandescent bulbs:
- 40 W ≈ 450 lm
- 60 W ≈ 800 lm
- 75 W ≈ 1100 lm
- 100 W ≈ 1600 lm
Efficacy — lumens per watt — is the real efficiency measure. Incandescent bulbs manage around 15 lm/W, compact fluorescents 60, and good LEDs 100 or more. Laboratory LEDs have exceeded 200.
Why lux depends on distance
Illuminance falls with the square of distance. Double the distance and you get a quarter of the light.
This is why a "lux" specification is meaningless without a distance. A security light rated "10,000 lux" might mean at one metre, where it is unremarkable, or at ten metres, where it would be extraordinary.
Typical illuminance levels:
- Direct sunlight: 32,000–100,000 lx
- Overcast day: 1,000–10,000 lx
- Office lighting: 300–500 lx
- Living room: 50–150 lx
- Full moon: about 0.25 lx
The range between full sunlight and moonlight spans about six orders of magnitude, which the human eye handles through adaptation.
Torch and headlight marketing
Two common inflations are worth recognising.
Emitter lumens versus out-the-front lumens. A manufacturer may quote the theoretical maximum of the LED die at ideal temperature and current. The finished torch, with reflector losses, lens losses and thermal throttling, might deliver 60–70% of that. Reputable brands quote ANSI FL1 figures measured from the assembled product.
Peak beam candela quoted as lumens. A tightly focused beam has very high candela but modest total flux. Quoting the candela figure with a lumen label produces an impressive and meaningless number.
The ANSI FL1 standard requires disclosure of output, beam distance, runtime, impact resistance and water resistance, measured in defined conditions. Products carrying FL1 figures are far more comparable.
Colour temperature and CRI
Two further specifications matter more than most buyers realise.
Correlated colour temperature, in kelvin, describes the warmth of the light — confusingly, higher numbers look cooler. 2700 K is warm white, 4000 K neutral, 6500 K daylight.
Colour rendering index (CRI) measures how faithfully colours appear under the light, on a scale to 100. Below 80 is poor, above 90 is good. Cheap LEDs often have CRI around 70, which makes skin tones and food look unpleasant regardless of brightness.
A 90+ CRI bulb at 800 lm will produce a far better-looking room than a 70 CRI bulb at 1200 lm.
Our illumination converter handles lux, foot-candle and phot; the luminous intensity converter covers candela and its historical predecessors.
Beam angle changes everything
Two bulbs with identical lumen ratings can produce completely different results, because lumens say nothing about where the light goes.
A 600-lumen spotlight with a 25 ° beam concentrates its output into a small area, producing high illuminance there and almost nothing elsewhere. A 600-lumen bulb with a 120 ° flood pattern spreads the same total output across a much wider area, producing lower illuminance everywhere.
For task lighting you generally want a narrower beam and higher illuminance on the work surface. For ambient lighting you want a wide beam and even coverage. Buying on lumens alone routinely produces a room that is simultaneously bright and badly lit.
Measuring light yourself
Phone lux meter apps use the ambient light sensor and are surprisingly serviceable for relative comparisons — checking whether one corner of a room is dimmer than another, or whether a desk meets a rough target.
They are much less reliable in absolute terms. Phone sensors have limited dynamic range, their spectral response differs from the human eye, and they sit behind a cover glass of unknown transmission. Treat readings as indicative, not calibrated.
Useful working targets: about 300–500 lx for general office work, 500–750 lx for detailed tasks such as sewing or fine assembly, and 100–200 lx for comfortable living space. Ageing eyes need substantially more — a person of 60 may need roughly twice the illuminance of a person of 20 for the same perceived clarity.