You can spec a candle’s scent load to a tenth of a percent, name the wick size, and hold the pour to a degree. But ask the same supplier how bright the finished flame is and the answer is a shrug. “Warm glow.” “Ambient.” “Close to a real flame.” None of those is a number, and none of them survives contact with a dining table that is two times too bright or two times too dim.
This is the gap. Light is the one property of a candle that nobody on the spec sheet agrees to measure. This post closes it. There is a unit for it, a law that predicts how far the light reaches, and a reason the LED replacement never looks the same under a person’s skin. All three fit on one page, and once you have them you stop guessing your flame count.
The unit is candela, and it is named for the candle
A single paraffin or soy taper emits a luminous intensity of roughly one candela. That is not a coincidence in the naming. The candela is the SI base unit of luminous intensity, and the word is Latin for candle. The modern definition pins it to a monochromatic 540 terahertz source, but the working fact a buyer needs is the one our own bench confirms: a normal wick flame lands at about 1 cd. Beeswax runs a little lower, a large paraffin pillar a little higher, and a wood-wick face splits its output into a wider, flatter beam, but one candela is the number to hold in your head.
Lumens are the wrong headline for a candle. Lumen measures total light leaving a source in all directions, which is exactly what a directional bulb is good at and what a candle is only roughly good at. A candle that gives off about 13 lumens in total sounds faint, yet at half a meter across a table it reads clearly. The lumen number flattens the geometry away. Candela keeps it. If your supplier will only quote lumens, ask for candela. If they will not, you are getting a number they do not understand.
One number predicts the whole room
A candle is close enough to a point source that the inverse-square law does the work for you. The illuminance on the table, in lux, is the intensity in candela divided by the distance in meters, squared:
E = I / d²
| Distance | Illuminance, 1 candle | What it reads like |
|---|---|---|
| 0.3 m | 11 lux | A face in close profile, strong shadow |
| 0.5 m | 4 lux | Typical dinner-table level from one flame |
| 1 m | 1 lux | Soft, moody, face falls into shadow |
| 2 m | 0.25 lux | Silhouette only |
Notice the falloff. Move a flame from the center of the table to the far edge and you lose roughly half the light at the seats, not a little of it. That is the whole reason a long banquet reads darker at the ends than anyone planned. The math that fixes it is the same one: because lux scales with the candle count when the distance is held, the number of flames you need at a given spot is the target lux multiplied by the distance squared. Want eight lux at a seat one meter out, you are around eight candles clustered within that meter. Want the same eight lux at two meters, you jump to thirty-two. That jump, not the number of candles on the table, is what separates a scene that feels lit from one that feels like a single flame is doing its best.
In our factory runs we measure the finished flame before we spec a retail set, not the wax. A client who ordered a twelve-flame LED dinner set told us the room felt “flat and greenish” even though the lux meter agreed with our spec. The lux was right. The lux was the wrong half of the question. The color was the other half.
Color is where the flame and the LED split
A candle flame sits near 1,000 to 1,900 kelvin, deep orange. The “warm white” LED that most flameless candles ship at is 2700 kelvin. Both are called warm. They are not the same warm. A 2700 K LED carries far more green and blue in its spectrum than a 1,000 K flame, and that extra blue is what turns skin toward green under the light and makes a face look lit by a lamp instead of by fire. We have watched this fail on ceremony sets: the candlelight looked correct on the wax, then the moment a person sat down the whole table read as a hospital hallway.
If you are matching an LED to a flame for a photo, a stage, or a retail display, the number to chase is color temperature, then color rendering, not brightness. A 2200 K LED with a high color rendering index will sit against a real flame without a seam. A 3000 K LED, even at the correct lux, will show its edge the second a camera opens its aperture. Specify the kelvin on the data sheet. “Warm white” is a marketing word, and it hides a 1,000 kelvin gap.
The efficacy number that explains everything else
A candle burns at roughly 0.1 grams per minute and releases about 80 watts of heat, yet only about 13 lumens of light. That is a luminous efficacy near 0.16 lumens per watt, on the order of a hundred times less efficient than an incandescent bulb. It is not a flaw to design around. It is the definition of what a candle is: a heat source that happens to make light. This is why a large flame in a small, unventilated space feels warm on the face before it feels bright, and why a tall pillar reads as glow more than illumination. When a buyer complains a candle “isn’t bright enough,” half the time the room simply needs more total heat in it, which more flames deliver along with more light. You cannot get one without the other. A 1,500-lumen LED can light the same table for a fraction of the heat, and that is exactly why it never feels like the candle it is replacing.
Four numbers to put on the spec sheet
Give the flame a candela figure, measured at the finished wick, not estimated from the wax. Give it a color temperature in kelvin, and if it is a flameless unit, hold it under about 2400 K to sit with real fire. Give it a flicker rate for any LED flame, because a steady LED glow is the first thing a trained eye rejects as fake even when the color is perfect. And give it a heat note, because a venue that is heat-limited will quietly trade candles for LED long before the light level is the problem.
None of this needs a lab. A phone with a lux app and a kelvin reference card covers the table check. What it does remove is the guesswork that a “warm, ambient, candle-like glow” was supposed to cover. Once the four numbers are on the sheet, the scene either meets the brief or it does not, and you can see which before the event, not after the client photographs the table.