Back to glossary

Term

dBFS, Headroom & Clipping

Three related concepts for digital audio levels: dBFS is the scale with 0 dBFS as the hard ceiling, headroom is the safety margin below the maximum, clipping is the hard cut-off when a signal exceeds it.

dBFS, Headroom & Clipping — explained in more detail

dBFS (Decibels relative to Full Scale) is the logarithmic scale for digital audio levels. Its reference is the numerical maximum a sample can take at the given bit depth — this ceiling is 0 dBFS. Real-world levels sit below it and are stated as negative values: -6 dBFS is half the amplitude, -20 dBFS a much quieter signal. Unlike dBSPL, dBFS is dimensionless and defined purely relative to the bit-maximum.

Headroom is the distance between the loudest point of a recording (peak) and 0 dBFS. It serves as a safety buffer: common mastering practice leaves -1 dBFS to -3 dBFS of peak headroom to absorb rounding errors from resampling, lossy encoders or mix-bus summing without later stages tipping into clipping.

Clipping happens the moment a signal would exceed 0 dBFS. Since no value above the maximum can be represented, the waveform is cut off hard — it goes flat at top and bottom. The result sounds rough, harsh and distorted. The original information is lost beyond recovery; clipping is destructive, not reversible.

Example / In practice

Clean dictation recordings typically sit between -12 dBFS and -3 dBFS peak. When they are fed to ASR models such as Whisper, clipping is especially damaging: distorted vowels look like broken phonemes to the model and lead to misreads. Software normalization to a -1 dBFS peak is therefore the standard move — hot level without digital overflow.

dBSPL measures acoustic sound pressure in the air and is not a digital concept. Peak is the highest instantaneous value, RMS is average energy over a window, LUFS is perceived loudness under broadcast standards. Headroom is normally stated against peak level, not against RMS or LUFS.

Entdecke mehr