Loudness targets · Explainer
True peak explained: why −1 dBTP
A sample-peak meter only sees the samples. The analog waveform they describe can swing higher between them. True peak estimates that higher value, and it's why every target on this site asks for −1 or −2 dBTP rather than 0 dBFS.
Samples aren't the waveform
Digital audio stores the level of the signal at regular instants: 48,000 times a second at 48 kHz. When it's played back, the converter reconstructs a smooth, continuous waveform through those points. That waveform doesn't have to peak exactly on a sample. It can crest between two of them.
A sample-peak meter reports the highest sample, in dBFS. A true-peak meter estimates the highest point of the reconstructed waveform, in dBTP (decibels true peak). It does this by oversampling, working out the in-between values at several times the original rate, as described in ITU-R BS.1770.
Inter-sample peaks
The difference between the two readings is the inter-sample peak. On gentle material it's tiny. On loud, heavily limited masters it can be a dB or more, because the limiter keeps pushing sample after sample right up to the ceiling, and the waveform between them overshoots.
A file with every sample at or below 0 dBFS can therefore describe a waveform that goes above full scale. Whether that becomes audible clipping depends on what happens next: a DAC with no headroom, a sample-rate converter, or a codec.
Why lossy encoding makes it matter
Most listeners don't hear your master. They hear a lossy encode of it, decoded on their device. Lossy encoders discard and reshape content to save data, and the decoded waveform can peak higher than the original, on top of any inter-sample peaks already there. The louder and denser the master, the more likely the decode overshoots.
A true-peak ceiling below 0 leaves room for all of that:
- −1 dBTP is the common ceiling: Spotify, Apple Music, YouTube, Tidal, Deezer, EBU R128 and Apple Podcasts.
- −2 dBTP gives more margin: Amazon Music, ATSC A/85 and Netflix. Spotify also suggests −2 dBTP for masters louder than −14 LUFS.
A master that meets −2 dBTP meets −1 dBTP too, so if you want one master for everything, the lower ceiling covers the whole table.
What to do about it
- Use a true-peak limiter (or a limiter with a true-peak or oversampling mode) as the last processor, set to your ceiling.
- Meter true peak, not just sample peak. A limiter set to −1 dBFS sample peak can still let −0.5 dBTP through.
- Check the whole song. The highest true peak is often a single transient in the loudest section. Maxima catch it.
- Loud masters need more room. If your integrated loudness is well above the service reference, consider −2 dBTP.
Sample rate and the phone
MixOutpost streams at 48 kHz. Sessions at 44.1, 88.2 or 96 kHz are converted to 48 kHz before streaming; 48 kHz sessions arrive bit-exact. In a converted session the phone measures the converted stream, so treat true peak as a close guide and confirm your final file with your limiter's own meter.
Checking true peak with MixOutpost
MixOutpost's loudness meter runs on your iPhone, fed by exactly what your Mac sends, and measures true peak alongside LUFS.
- 01
Stream your master bus
Insert the MixOutpost plugin last on the master, after the limiter, and press Start Sharing. Pair your iPhone once with a QR code.
- 02
Open Loudness
Alongside momentary, short-term and integrated LUFS and LRA, you'll see the true-peak maximum since the last reset. Compare it with your target's ceiling in the table below.
- 03
Reset and play the whole song
The true-peak maximum holds the highest value until you reset. It measures what the Mac sends, before Listen as, Dim, Mute or phone volume.
True-peak ceilings across targets
| Target | Integrated | True peak |
|---|---|---|
| Spotify | −14 LUFS | −1 dBTP |
| Apple Music | −16 LUFS | −1 dBTP |
| YouTube | −14 LUFS | −1 dBTP |
| Amazon Music | −14 LUFS | −2 dBTP |
| Tidal | −14 LUFS | −1 dBTP |
| Deezer | −15 LUFS | −1 dBTP |
| EBU R128 | −23 LUFS | −1 dBTP |
| ATSC A/85 | −24 LKFS | −2 dBTP |
| Netflix | −27 LKFS | −2 dBTP |
| Apple Podcasts | −16 LUFS | −1 dBTP |
More on each target on the loudness targets page.
Questions
What's the difference between dBFS and dBTP?
dBFS sample peak is the highest sample value in the file. dBTP true peak is an estimate of the highest value of the continuous waveform those samples represent, found by oversampling. True peak is always equal to or higher than sample peak.
Can a file that never hits 0 dBFS still clip?
Yes. If the reconstructed waveform peaks above full scale between samples, a DAC, a sample-rate converter or a lossy decoder can clip it. That's an inter-sample peak, and a true-peak meter shows it where a sample-peak meter doesn't.
Why do some specs ask for −2 dBTP instead of −1?
More headroom for processing after you. Loud, dense masters and audio that will be encoded or processed again downstream are more likely to overshoot. Amazon Music, ATSC A/85 and Netflix use −2 dBTP.
Does true-peak limiting change the sound?
It can. A true-peak limiter has to catch overshoots the waveform hasn't reached yet at the sample level, so it may reduce more than a sample-peak limiter at the same ceiling. On most material the difference is small; on dense masters, listen for it.
Hear your mix on your phone
Free Mac plugin (AU, VST3) and iPhone app. Lossless, bit-exact and encrypted.