True Peak Explained: Why Masters Clip on Spotify (-1 dBTP)

True peak (dBTP) measures the loudest point of the reconstructed analog waveform — including peaks that land between digital samples. A master can read 0 dBFS on a sample meter and still clip once Spotify or Apple Music re-encodes it. Keeping true peak at or below -1.0 dBTP is the streaming-safe ceiling.

What is the difference between true peak and sample peak?

A digital audio file isn't a waveform — it's a list of measurement points taken thousands of times per second. Sample peak (what a normal dBFS meter shows) is simply the highest of those points. But when your file is played back, the DAC reconstructs a smooth, continuous curve through the points, and that curve can swing above them.

Picture the crest of a wave with one sample just before the top and one just after, both sitting at -0.1 dBFS. The real waveform arcs over the gap between them — higher than either sample. True peak, defined in ITU-R BS.1770, catches this by oversampling the signal (measuring the curve at several points between each sample) to estimate where the reconstructed waveform actually tops out. That's why a track's dBTP reading is always equal to or higher than its sample peak — sometimes by 1 dB or more on loud, heavily limited material.

How can a 0 dBFS file still clip? Inter-sample peaks

Those between-the-dots overshoots are called inter-sample peaks, and they're why "my file never touches 0 dBFS" doesn't mean "my file never clips." A master limited hard to 0 dBFS can reconstruct to well over 0 dBTP — the samples are legal, but the waveform they describe isn't.

Two things then push those peaks into audible trouble. First, playback hardware: DACs and cheap output stages have no headroom above full scale, so the reconstructed overshoot gets flattened. Second — and far more common for streaming artists — lossy encoding, which is the next section. The frustrating part is that most stock limiters and clippers only watch sample values, so they'll happily certify a master as "peaking at -0.1 dB" while its true peak sails past zero. If your master sounds subtly crunchy on streaming but clean in your DAW, inter-sample clipping is a prime suspect.

Why does lossy encoding (Ogg and AAC) add peaks?

Streaming platforms don't play your WAV. Spotify transcodes to Ogg Vorbis and AAC; Apple Music serves AAC; YouTube re-encodes everything. A lossy codec doesn't preserve your waveform — it stores a perceptual approximation and rebuilds something close-but-not-identical on decode. That rebuilt waveform has slightly different peaks, and on loud material they're usually slightly higher.

If your master already sits at 0 dBFS, there's nowhere for those new peaks to go except into clipping. This is exactly why Spotify's own mastering guidance says to keep true peak below -1 dBTP "best for lossy formats," and why Apple's Apple Digital Masters program checks every AAC encode for clipping and pushes engineers to leave headroom (both as of July 2026). The -1.0 dBTP ceiling isn't superstition — it's a buffer sized for what encoders actually do to peaks.

What true peak should I master to? The -1.0 dBTP rule

For almost everything, -1.0 dBTP. Here's the platform-by-platform picture:

Verified July 2026 — we re-check these quarterly.

Notice Spotify's second clause: if you master hotter than -14 LUFS, they ask for -2 dBTP, because normalization turning a hot master down doesn't undo the distortion baked in during encoding. True peak and loudness are separate targets — our LUFS streaming targets guide covers the loudness half.

True-peak limiting vs plain limiting: what's the difference?

A plain (sample-peak) limiter constrains the dots. A true-peak limiter oversamples its own output, estimates the reconstructed waveform between the dots, and keeps that under the ceiling. The trade-off is honest: true-peak limiting gives up a fraction of a dB of loudness compared to slamming samples against zero. That fraction is inaudible; encoder clipping isn't.

This matters because the usual reason people push peaks to zero is chasing loudness — and on normalized platforms that race is already lost. Spotify turns loud masters down to its playback level anyway, so a 0 dBFS master buys you nothing except the clipping. If your real worry is that your track sounds quiet next to commercial releases, that's a density and tonal-balance problem, not a peak problem — we break that down in why your song sounds quieter on Spotify. TrackGleam's mastering chain uses a true-peak ceiling of -1.0 dBTP and measures the actual dBTP of every master it renders, so you're not trusting a setting — you're reading a measurement.

Do AI-generated tracks have true-peak problems?

Some do. We measured 12 AI-generated music exports from our test library, and 3 of 12 exceeded the -1.0 dBTP true-peak ceiling — files that would be pushed into extra distortion by streaming transcodes. The loudness side was scattered too: median -15.2 LUFS integrated, with 8 of 12 sitting below the -14 LUFS streaming level. In other words, AI platforms hand you exports that are quiet and occasionally over-peaked at the same time — the worst combination, since simply turning them up would push more peaks over the line.

If you're releasing AI-generated tracks, mastering isn't optional polish — it's the step that fixes both numbers at once. Our guide to mastering AI-generated songs walks through the whole workflow, including the artifact-cleanup presets built for AI audio.

How do I check my track's dBTP for free?

Drop your track into TrackGleam. Every master it produces reports integrated LUFS (ITU-R BS.1770-4, gated) and dBTP true peak — measured on the finished file, not assumed from limiter settings. Processing runs entirely in your browser: no upload, no account, no watermark, and the free master targets -14 LUFS with a -1.0 dBTP ceiling out of the box. If you just want the numbers, run a master and read the report; if the numbers are already fine, you've spent two minutes confirming your release is encoder-safe.

Master a track free — no signup, nothing uploads