Dimmer Curves: Linear, Square Law and S-Curve Compared

Push the fader to the middle and you expect half brightness. You rarely get it — the eye reads it as much brighter than half. Nothing is broken; human vision simply does not respond linearly to light. Dimmer curves exist to close that gap.

The eye is not linear

Doubling a fixture’s light output does not double perceived brightness. To cope with an enormous range of light levels, human vision compresses its response: in darkness we notice tiny changes, in brightness we struggle to see large ones.

The practical result: map DMX values linearly to output power and the light rushes up at the bottom of the fader, then appears to barely change at the top.

Common curves

Curve	Behaviour	Where it is used
Linear	DMX value maps straight to output power	Measurement, LED strip drivers, non-lighting uses
Square law	Output = (value/255)² — linearises perceived brightness	General stage use, most LED fixtures
S-curve	Eases at both ends, faster through the middle	Theatrical fades where the top and tail should be gentle
Tungsten emulation	Mimics the thermal lag and warm shift of an incandescent lamp	Scenes mixing LED fixtures with conventional lamps

Why square law dominates

Because the eye roughly follows a power law. Scale output by the square of the DMX value and the middle of the fader genuinely feels like half:

DMX 128  ->  linear:      50% power   (looks much brighter than half)
DMX 128  ->  square law:  25% power   (looks roughly half)

LED versus conventional lamps

A tungsten filament takes time to heat and cool. That lag provides accidental but welcome smoothing: even abrupt DMX changes look gentle in the light. It also shifts colour temperature downwards as it dims, warming the light — a familiar and much-loved effect in theatre.

An LED has no such inertia; the light changes the instant the value does. That creates two problems: fades can look stepped, and dimming stays “cold” because colour temperature does not move. Good LED fixtures solve both in software, applying a curve and offering an optional tungsten mode that warms as it dims.

Stepping at the bottom end

The problem is most visible just as the light comes off zero. On an 8-bit dimmer channel the difference between values 1, 2 and 3 produces visible jumps. Three fixes:

• Use a 16-bit dimmer channel — 65,536 steps instead of 256.
• Raise the fixture’s PWM resolution and frequency, if the menu allows it.
• On very slow fades, avoid the bottom end entirely and start the cue around 3–5%.

PWM frequency when cameras are involved

A dimmer curve fixes perceived brightness; it does nothing about banding on camera. That comes from the fixture’s PWM frequency. If you are shooting, raise the PWM frequency in the fixture menu (above 20 kHz where possible) and test against the camera shutter. Budget fixtures often do not offer the setting at all.

Practical advice

1. Set every fixture to the same curve; mixed curves make lights in one scene come up at different rates.
2. Apply the curve in one layer only (console or fixture).
3. When LEDs share a scene with conventional lamps, try tungsten emulation on the LED side.
4. Use 16 bit for dimmer on long fades.
5. After any change, actually black out the room and watch the fade; the numbers can look right while the eye disagrees.

Related articles

• 16-Bit DMX: Coarse and Fine Channels