−18 dBFS, −12 dBFS, −20 dBFS: Reference Levels, dBu and dBV
Articles in English
What you'll learn: where −18 dBFS and −20 dBFS come from, what −12 dBFS really is, how dBFS maps to analog dBu and dBV, and how to set recording levels with these numbers.
Recording guides keep repeating the same numbers: record at −18 dBFS, keep peaks under −12 dBFS, align at −20 dBFS. Some are standards, one is a habit, and all of them make more sense once you see how digital levels connect to the analog world.
1. Why Reference Levels Exist
Analog equipment has a nominal operating level, read as 0 VU on a meter, and saturates gradually above it. Digital audio has a hard ceiling: nothing exists above 0 dBFS, and a signal that tries to go higher is clipped. An alignment level decides where the analog nominal level lands on the digital scale. Everything above it is headroom for peaks.
2. −18 dBFS and −20 dBFS: the Two Alignments
| Convention | Alignment tone | Analog level | 0 dBFS equals |
|---|---|---|---|
| EBU R68 (Europe) | −18 dBFS | 0 dBu | +18 dBu |
| SMPTE RP 155 (US) | −20 dBFS | +4 dBu | +24 dBu |
EBU R68 places the alignment level at −18 dBFS, which European broadcasters tie to 0 dBu on the analog side: 18 dB of headroom. SMPTE RP 155 places the reference at −20 dBFS, usually tied to +4 dBu in US studios: 20 dB of headroom. The analog side of these alignments is a matter of how each facility calibrates its converters, so check the maximum output level in your interface's datasheet rather than assuming one.
3. −12 dBFS: a Habit, Not a Standard
No standard defines −12 dBFS: EBU R68 and SMPTE RP 155 define −18 and −20 dBFS. The value appears in recording advice as a peak target, and the headroom it keeps is cheap at 24 bits. The format's theoretical dynamic range is about 144 dB, and high-end audio converters are specified above 110 dB, so a signal peaking 12 dB below full scale still sits roughly 100 dB above the converter's noise floor.
At 16 bits the margin is smaller: about 96 dB of theoretical range.
4. Peak or RMS?
These numbers are read on different meters. Alignment levels refer to a steady sine tone, for which a sample peak meter and an RMS meter following the AES17 convention agree: AES17 defines a full-scale sine as 0 dBFS. Some RMS meters do not follow that convention and show a sine 3 dB lower, so a −18 dBFS tone can read −21 dB on them. Peak targets such as −12 dBFS are read on a peak meter.
5. dBFS, dBu, dBV and dB SPL
| Unit | Reference (0 dB) | Used for |
|---|---|---|
| dBFS | Digital full scale (the largest sample value) | Digital meters, files, DAWs |
| dBu | 0.775 V RMS (1 mW into 600 Ω) | Professional analog line level |
| dBV | 1 V RMS | Consumer line level, mic sensitivity |
| dB SPL | 20 µPa of sound pressure | Acoustic level in the room |
dBu = dBV + 2.2
dBu = dBFS + (level in dBu at 0 dBFS)
+4 dBu = 1.228 V RMS (professional line level)
−10 dBV = 0.316 V RMS (consumer line level, ≈ −7.8 dBu)
Example: interface with 0 dBFS = +18 dBu
−18 dBFS → 0 dBu
−12 dBFS → +6 dBuThe professional and consumer line levels are about 11.8 dB apart, which is why plugging consumer gear into a +4 dBu input sounds weak and the reverse can overload.
dBFS and dB SPL are linked in the same way, by a calibration offset instead of a converter's maximum level. The dBFS to dB SPL converter uses the common convention 0 dBFS = 94 dB SPL, with a correction field for your own offset.
Delivery loudness is a separate question, measured in LUFS rather than dBFS: see LUFS vs dBFS.
Know the Level Before You Record
Enter the loudest sound level you expect in dB SPL and the converter gives the dBFS it will reach, with your microphone's sensitivity correction.