True Peak Safety

True peak safety is the practice of leaving enough margin so a mastered file does not clip after reconstruction, codec conversion, or streaming playback.

Sample peaks only show the values stored in the digital file. Playback systems and encoders reconstruct a continuous waveform between those samples. That reconstructed waveform can rise above the stored sample peaks; the part that pokes above is called an inter-sample peak (ISP). A master that looks safe at sample level can still distort after conversion.

LOUDNESS · TRUE PEAKIDLE

Inter-sample peaks — why a master clips on playback

Scroll to load

The dots are samples of a sine wave; the curve shows its analytic waveform. A crest can fall between samples, above every stored value. The gap depends on frequency and phase, so it does not increase smoothly toward Nyquist. Compare the analytic peak with the oversampled meter reading. Raising the sample peak toward 0 dBFS can leave an inter-sample over that clips a converter or processor without sufficient headroom.

Sample peak

-0.3 dBFS

Frequency (×Nyquist)

0.4

Why Sample Peak Is Not Enough

A digital audio file is a series of points, not a complete curve. When a DAC (digital-to-analog converter) or encoder turns those points back into sound, it reconstructs the curve implied by the samples. The curve can rise above the points.

True-peak measurement estimates this by oversampling the signal before checking peaks. How far the two readings differ depends on the material:

Signal	Sample peak (dBFS)	True peak (dBTP)
Sine at a quarter of the sample rate, 45° phase	-3.01	-0.09
Square wave, amplitude 0.5	-6.02	-4.11
mixed-stems.wav	-1.1	+0.2
drum.wav	-0.7	+1.4

The last two rows are files that sit below 0 at sample level and above 0 once reconstructed. drum.wav looks safe at -0.7 and reconstructs about 2 dB higher.

This matters most when:

Situation	Risk
Lossy encoding	MP3, AAC, or Opus conversion can create new peaks.
Loud limiting	Dense masters leave less room for reconstruction overshoot.
Streaming normalization	Playback pipelines may transcode or process the file.
Consumer devices	Some playback chains handle near-0 dBFS material poorly.

Practical Ceiling

The ceiling is the highest level the limiter will let through. A common release ceiling is around -1 dBTP — decibels true peak, the same decibel scale as dBFS but measured on the reconstructed waveform. Some workflows use -2 dBTP for extra codec safety; others use a tighter ceiling when the destination is known and controlled. The engine's club and cd delivery targets use -0.3 dBTP together with -9 LUFS, which trades codec safety for density; streaming, broadcast, and podcast targets use -1.

The ceiling is not a loudness target. It is a safety margin. You can have a quiet file with unsafe peaks or a loud file that stays within true-peak limits.

METERS · LOUDNESSIDLE

Loudness metering — LUFS, true-peak, and range

Scroll to load

The bar tracks momentary loudness as the clip plays; the panel is the loudness over time. Integrated LUFS is the single overall number, true-peak estimates peaks between samples, and LRA captures how much the loudness moves. Switch the window to compare the fast momentary meter with the smoother short-term one — each reports only once its window has filled, so the short-term contour starts three seconds into the clip.

Window

Limiter Lookahead

A true-peak limiter needs time to react before the peak arrives. Lookahead gives the limiter that time. Too little lookahead can miss fast peaks. Too much lookahead can increase latency and may change transient feel.

In offline mastering, a few milliseconds of lookahead is usually acceptable because playback latency does not matter. In live or interactive systems, latency matters more.

The limiter's defaults:

Setting	Default
ceilingDb	-1
lookaheadMs	1.0
releaseMs	50
oversampleFactor	4

Measure it in libsonare

The true-peak meter and the limiter both oversample 4x by default and accept 1, 2, 4, 8 or 16. Limiting to a ceiling is a chain stage, maximizer.truePeakLimiter.

Browser

import { init, meteringPeakDb, meteringTruePeakDb, masteringChain } from '@libraz/libsonare';

await init();

const samplePeak = meteringPeakDb(samples, sampleRate);
const truePeak = meteringTruePeakDb(samples, sampleRate);      // 4x oversampling
const finer = meteringTruePeakDb(samples, sampleRate, 8);      // 1, 2, 4, 8 or 16

const limited = masteringChain(samples, sampleRate, {
  maximizer: { truePeakLimiter: { ceilingDb: -1, lookaheadMs: 1, oversampleFactor: 4 } },
});
// limited.samples holds the result; limited.outputTruePeakDbtp reads -1.0 for drum.wav

Python

import libsonare as sonare

audio = sonare.Audio.from_file("song.wav")
sample_peak = audio.peak_db()
true_peak = audio.true_peak_db()                      # oversample_factor=4
finer = audio.true_peak_db(oversample_factor=8)

limited = sonare.mastering_chain(
    audio.data, audio.sample_rate,
    {"maximizer": {"truePeakLimiter": {"ceilingDb": -1, "lookaheadMs": 1, "oversampleFactor": 4}}},
)

CLI

sonare meter song.wav   # native CLI only: sample peak and true peak (--oversample sets the factor)
sonare mastering-processor song.wav --processor maximizer.truePeakLimiter \
  --params ceilingDb=-1,lookaheadMs=1 -o limited.wav

The mastering demo's Studio mode exposes the same ceiling and lookahead controls.

Implementation notes

The safety check uses the same principle as true-peak metering: estimate the reconstructed waveform by oversampling, then evaluate the highest sample on that denser grid. A limiter that only clamps original samples can still leave inter-sample overs, so the guard has to observe the oversampled path, not just the input buffer.

Lookahead and release are coupled.

Lookahead gives the limiter a chance to begin attenuation before the peak. Release determines how quickly it returns to unity gain.

Very short release can produce modulation. Very long release can dull transients and reduce loudness headroom. The practical setting is a compromise between codec safety, transient preservation, and audible gain movement.

Related: True Peak, LUFS, Loudness Matching