A true peak limit is needed because audio can exceed 0 dBFS after it leaves your DAW—even if your sample peak meter never hits 0. Those “between-the-samples” overs can trigger distortion in real playback chains and in encoding/processing steps, so a true peak ceiling is a practical safety margin, not a mastering superstition.
The problem true peak limiting solves (in plain terms)
Digital audio is stored as snapshots (samples). Your DAW’s standard peak meter usually reports the highest snapshot value (“sample peak”). But listeners don’t hear snapshots—they hear a reconstructed waveform produced by a digital-to-analog converter (DAC). During reconstruction, the curve between samples can rise higher than any individual sample. That rise is an intersample peak (often discussed under “true peak”).
If your loudest sample sits at -0.1 dBFS, you might assume you’re safe. But reconstruction can push the analog waveform past 0 dBFS-equivalent. Once you cross that ceiling in a real device or processing stage, the result is clipping or gritty transient distortion—sometimes subtle, sometimes obvious, and often inconsistent across devices.
Why sample peaks miss it
A sample peak meter only checks the values that exist in the file. The “actual” peak of the continuous waveform can fall between those stored points. The only way to estimate that continuous peak in software is to oversample (interpolate) the signal and look for peaks in the higher-rate representation. That’s what true peak meters and true peak limiters are designed around: they try to predict what a DAC (or subsequent processing) will reconstruct.
Different true peak meters can disagree a little because oversampling methods, filters, and tolerances differ. The key point still holds: sample peak = what’s in the file; true peak = what the waveform can become in playback/processing.
Real-world playback chains are not uniform
Even if your studio playback sounds fine, consumers aren’t listening through your converter. Phones, TVs, soundbars, car stereos, Bluetooth devices, and inexpensive DACs can behave differently near full scale. Some handle overs gracefully; others clip earlier or have less headroom in internal stages. True peak limiting is a way to avoid “it clips on my friend’s phone but not here.”
This is why true peak issues are often reported as:
- “It distorts only on certain speakers.”
- “It’s clean in the DAW, crunchy on streaming.”
- “The master is fine until it gets uploaded/encoded.”
Encoding can create new peaks (even if you were safe before)
Lossy codecs (AAC, MP3, Opus, etc.) don’t preserve the waveform exactly. They approximate it. During encode/decode, tiny changes in phase and transient shape can produce level overs relative to the original PCM. That means a file that never exceeded 0 dBFS in the DAW can decode to something that effectively does—especially with very dense, bright, transient-heavy material.
True peak headroom helps because it gives the codec “room to move” without smashing into the ceiling. In practice, this is one reason you’ll see platform guidance that includes a true peak maximum, not just loudness.
Platform processing can also push levels around
Many distribution chains do more than just “host your file”:
- loudness normalization,
- transcoding to multiple codecs/bitrates,
- device-specific playback paths,
- ad insertion or stitching,
- automatic level management in broadcast-like contexts.
Even if normalization primarily turns audio down, intermediate steps can still create peaks. And in some workflows (especially ad tech and broadcast chains), audio can be measured, gated, limited, or converted multiple times. A true peak ceiling keeps your master robust across those steps.
Spotify’s ad guidance, for example, explicitly includes a true peak maximum (and it’s not the same as “keep your samples below 0”). (adshelp.spotify.com)
Why “0 dBFS sample peak” is a fragile target
Hitting 0 dBFS sample peak is like packing a suitcase to the exact millimeter: it might close in your room, then fail at the airport when the zipper flexes.
A master that kisses 0.0 dBFS on sample peaks has no tolerance for:
- reconstruction overs in consumer DACs,
- codec-induced overs,
- slight resampling differences (44.1 ↔ 48 kHz conversions),
- downstream processing that changes transient shape.
This is why “it doesn’t clip in my DAW” is not a reliable test. Your DAW is typically reporting the samples. Playback and distribution often behave closer to the reconstructed waveform reality.
What a true peak limiter is actually doing
A true peak limiter is usually a limiter that:
- Oversamples internally (commonly 2x, 4x, 8x, sometimes more),
- Applies limiting while “seeing” the higher-rate peaks,
- Produces output that, once reconstructed, is less likely to exceed the ceiling you set.
The ceiling is expressed in dBTP (decibels true peak). If you set -1.0 dBTP, you’re not saying “samples will never exceed -1.0 dBFS.” You’re saying “the reconstructed peak estimate should stay below -1.0 dBTP.”
Why common ceilings are negative numbers (like -1.0 dBTP)
Because the goal is headroom. In much of modern delivery, -1.0 dBTP is used as a practical compromise: enough margin to prevent many real-world overs, without forcing the master to be audibly quieter in any meaningful way.
Some pipelines are stricter. For example, certain ad delivery specs call for more margin (e.g., -2 dBTP) because ads can be stitched, transcoded, and played in a wide range of contexts where a clean ceiling matters more than maximum level. (adshelp.spotify.com)
When true peak limiting is most “needed”
True peak limiting provides the most value when at least one of these is true:
- Your audio is destined for streaming or web distribution where codec conversion is guaranteed.
- Your mix is transient-rich (sharp drums, clicks, aggressive consonants, bright percussion). These are more likely to produce intersample overs.
- Your master runs very close to full scale (hot limiting, dense material).
- You’re delivering to a spec (broadcast, ads, or any platform guidance that mentions dBTP).
- You have to be reliable across devices (music releases, podcasts, ads, and video content meant for broad playback environments).
In other words: if the audio will leave your controlled playback environment, true peak management becomes a reliability step, not an optional tweak.
When true peak limiting is less critical (but still useful)
If your peak ceiling is already conservative (say, your loudest sample peaks are around -3 dBFS and you aren’t pushing loudness hard), you may have enough incidental headroom that true peaks won’t be a practical problem. But many modern masters—especially ones that get close to 0 dBFS—don’t have that cushion.
Even then, using true peak metering is still useful as a verification step. You don’t necessarily need heavy limiting; you need certainty that your ceiling will hold up after distribution.
True peak is also about avoiding “mystery distortion”
One of the most frustrating outcomes in audio delivery is distortion that:
- doesn’t show up in your meters,
- doesn’t happen in your studio playback,
- appears only after upload or only on some devices.
True peak limiting reduces the odds of that scenario. It doesn’t guarantee perfection—no single tool can account for every possible playback chain—but it tackles a common, measurable failure mode: reconstructed/codec-induced overs.
The “standard” angle (why dBTP exists at all)
True peak measurement is not a plugin-maker invention; it’s baked into widely used loudness and metering standards. The ITU’s BS.1770 recommendation explicitly covers algorithms for loudness and true-peak level, which is why dBTP appears in professional delivery specifications. (ITU)
Practical takeaway: what you’re buying with a true peak limit
A true peak limit buys you:
- fewer codec-related overs,
- fewer device-dependent clipping surprises,
- easier compliance with platform specs,
- cleaner transients in real playback.
It’s less about “making it louder” and more about “making it survive the trip.”
why does this matter
Most listeners won’t hear your DAW session—they’ll hear a streamed, transcoded version on unpredictable hardware. A true peak limit is a small safeguard that prevents avoidable distortion where it actually counts.
sources
- ITU-R BS.1770 recommendation page (true-peak measurement context). (ITU)
- Production Advice: DSP overshoot, intersample peaks, and true peak limiting (practical explanation). (Production Advice)
- Spotify Ads help: audio ad loudness and true peak limits (example of real delivery specs). (adshelp.spotify.com)
