Home › Guides › Monitor calibration

How to calibrate your studio monitors

Calibrating decides what 0 dB on your volume control means in the room: a known level, set with pink noise and an SPL meter, one loudspeaker at a time. It takes ten minutes and an inexpensive meter.

By Jean-Pierre Baroni, acoustic designer · 11 September 2026 · Français

What calibration is for

A volume control is just a number until it is tied to the room. Calibration ties it: at 0 dB, pink noise at a defined digital level produces a defined sound pressure level at your ears. From then on your reference is a real place you can return to, the same every day and comparable with other rooms.

That buys three things: decisions that stay consistent from one session to the next, levels comparable with the film, broadcast and mastering standards, and the certainty that your normal level is not quietly creeping towards 95 dB. It is not required for loudness compensation, which only needs a truthful zero, as our loudness compensation guide explains.

Choose a reference

The ITU figure is the general rule 85 − 10 log n dB(A) per loudspeaker, n being the number of channels: 82 dB(A) each for a stereo pair, 78 dB(A) for five channels, which is where that older number comes from. K-20 and K-14 are two steps of Bob Katz's K-System: the same target, with a different amount of headroom above it.

EBU R 128 is not in the table on purpose. It normalises programme loudness to −23 LUFS and says nothing about how loud you listen.

All these targets are loud for a long session in a small room. Calibrating to one does not mean working at it all day: it fixes the reference, and loudness compensation lets you work lower without your bass decisions drifting.

What you need

• An SPL meter with A and C weighting and a slow response. A basic model is enough; a phone app will do for a rough check.
• Wideband pink noise at the reference level. Most DAWs can generate it, and so do monitor controllers.
• Your monitors in their normal position, and a fixed listening position.

Measure one loudspeaker at a time

Every one of these targets is defined per loudspeaker, not for the pair playing together. Two speakers reproducing the same noise read several decibels higher at the listening position, so a calibration made with both playing is wrong by that much. Play the noise through one side, measure, then do the other.

The procedure

1. Turn your monitors down before generating any noise.
2. Put the meter at the listening position, at ear height, pointing at the speakers. Set it to a slow response and to the weighting of your reference: A for ITU, C for SMPTE and the K-System.
3. Set your monitor volume to 0 dB and play the reference pink noise through the left speaker only.
4. Adjust the level on the monitors, their input sensitivity or your interface's output until the meter reads the target.
5. Do the same for the right speaker. The two should end within about a decibel of each other; a larger gap is a placement or gain problem worth finding before you go on.
6. Repeat for any other pair you switch to, to the same target. Headphones cannot be measured this way: match them by ear against the speakers.

Band-limited noise: for comparing, not for the absolute level

Pink noise filtered to a band, typically 500 Hz to 2 kHz, gives a steadier reading, far less sensitive to room modes, which helps when you compare one speaker with the other. The references above are set with wideband pink noise, so use the unfiltered signal while you set the absolute level.

Common mistakes

• Measuring with both speakers playing.
• A fast meter response or the wrong weighting.
• Holding the meter at the desk instead of the listening position, at ear height.
• Calibrating with a fader you will move later.
• Setting the absolute level with band-limited noise.