How to set a true-peak ceiling (and why −1 dBTP)
Once you know true peak can exceed sample peak (see True peak vs sample peak), the practical question is simple: what ceiling do I set on my limiter? For most masters, the answer is −1 dBTP. Here's why, and how.
Why not 0 dBFS
Setting a ceiling at 0 dBFS — or −0.1 dBFS on a sample meter — leaves no room for two things that happen after your limiter:
- Inter-sample overshoot on playback: the reconstructed waveform can rise above your last sample.
- Lossy encoding (MP3, AAC): the encoder reshapes the waveform and can push peaks higher still.
A little headroom absorbs both. That headroom is what a true-peak ceiling buys you.
Why −1 dBTP
−1 dBTP is the widely-used default, and it's not arbitrary:
- It reserves 1 dB for inter-sample peaks and encoder overshoot — enough for typical program material.
- It matches common streaming guidance: AES TD1008 recommends a maximum true peak no higher than −1 dBTP at the lossy codec input; Spotify asks for true peak below −1 dBTP. Apple asks for at least 1 dB of headroom, without specifying dBTP or true peak.
Two situations to nudge lower:
- Heavily-limited, dense masters push more energy into inter-sample peaks — −2 dBTP is a safer bet before aggressive lossy encoding. Spotify explicitly asks for true peak below −2 dBTP when your master is louder than −14 LUFS integrated.
- Spoken-word / podcast delivery specs sometimes ask for −2 dBTP outright.
If you're unsure, −1 dBTP is the safe default; −2 dBTP is the conservative one. There's rarely a reason to go above −1.
How to set it
- Put a true-peak limiter last in your master chain (a plain sample-peak limiter won't protect you — it can't see inter-sample peaks).
- Set the output ceiling to −1.0 dBTP as a starting point. For Spotify, keep measured true peak below −1 dBTP, or below −2 dBTP if your master is louder than −14 LUFS integrated.
- Confirm the limiter is in true-peak mode, not sample-peak mode (on some plugins this is a switch; on a dedicated true-peak limiter it's the default).
- Drive loudness with input gain into the limiter until it reads right on your LUFS check (see LUFS targets for streaming) — the ceiling holds the peaks while you set level.
Common mistakes
- Ceiling set at −0.1 dBFS sample and assuming you're safe — that's not a true-peak ceiling.
- True-peak metering off, so overs never show.
- Chasing loudness by lowering the ceiling — the ceiling controls peaks, input/drive controls loudness. They're different knobs.
Takeaways
- Set a true-peak ceiling, last in the chain, at −1 dBTP for most work; −2 dBTP for dense masters or podcast specs. Spotify asks for measured true peak below −2 dBTP when the master is louder than −14 LUFS integrated.
- The ceiling protects against inter-sample and encoder overshoot; input/drive sets loudness.
- Verify true-peak mode is actually on.
Part of the Loudness and true peak guide.
Sources
- Spotify: −14 LUFS integrated; true peak below −1 dBTP, or below −2 dBTP for masters louder than −14 LUFS. Spotify for Artists — Loudness normalization. Checked 2026-09-16.
- Apple: at least 1 dB of headroom, with no numerical LUFS reference or dBTP / true-peak recommendation in the reviewed guidance. Apple Digital Masters white paper / Apple Digital Masters. Checked 2026-09-16.
- −16 LUFS for track normalization comes from AES TD1008, not Apple. AESTD1008.1.21-9 (2021-09-24, supersedes TD1004), Table 1; maximum true peak ≤ −1 dBTP at the lossy codec input. These are recommendations for streaming-service normalization, not artist mastering targets. AES TD1008 PDF / AES document page. Checked 2026-09-16.
- YouTube / Amazon Music / Tidal: no official public numerical reference found in the reviewed sources. Treat levels reported for these services as measured / reported values only, not published platform specifications. Official-source search: YouTube / Google support, Amazon Music / Amazon Music for Artists, Tidal support / Tidal for Artists. Checked 2026-09-16.