Recognize sample clipping
In fixed-point digital audio, full scale is the maximum representable amplitude. Pushing a signal beyond it can flatten peaks and create distortion. A few samples at full scale are a warning to inspect, not automatic proof of an audible fault; deliberate clipped production and accidental converter overload are different situations.
Lowering a clipped file after the event creates headroom around the already distorted shape. Recovery tools may estimate a smoother contour, but they cannot know the original waveform with certainty. The best fix is usually at the unclipped mix or recording stage.
Understand the space between samples
A digital meter that reports only the largest stored sample can miss a higher continuous waveform reconstructed between sample points. These inter-sample events are why the ITU true-peak method estimates the signal at a higher effective sample rate.
True peak is reported in dBTP. It is an estimate that depends on a defined reconstruction model, not a direct measurement of every possible digital-to-analog converter. Still, it is more useful than sample peak alone when preparing audio for unknown downstream systems.
Leave room for later processing
Codec encoding, equalization and sample-rate conversion can alter peak shape. A master sitting extremely close to 0 dBFS may therefore overshoot later even when its stored samples do not clip. The appropriate margin depends on destination, loudness and the processing still ahead.
For Spotify, current public guidance recommends no higher than -1 dBTP, or -2 dBTP when the master is louder than -14 LUFS. Other destinations can differ. Verify the current specification instead of applying one number everywhere.
Measure and listen together
Inspect sample peak, true peak, loudness and clipped-sample counts, then listen around every flagged event. A snare transient that briefly triggers a warning may be less damaging than sustained flat-topping on a vocal.
Compare any limiter change at matched loudness. More headroom is useful when it prevents downstream distortion; it is not an end in itself. Preserve the version whose transient shape and tone best support the track while meeting delivery constraints.
METHOD
How this guide was prepared
Definitions follow ITU-R BS.1770 and the AES loudness project. Platform-specific margins are identified as current platform guidance rather than universal standards.
This is general educational information, not a guarantee of platform acceptance, pressing approval, authorship, or detector outcome. Requirements and services can change; check the receiving party before delivery.
PRIMARY SOURCES