In this lesson
True peak bites after encoding
A master can read -0.1 dBFS and still clip on a phone. Why the wave between samples and the lossy encode both peak higher, and which ceiling to set.
Updated · 6 min read · Lesson 40 of 44 in Mixing & Mastering
In short
- A sample-peak meter reads the samples, but the speaker plays the continuous wave between them, which can crest higher.
- Lossy encoding changes the waveform again, so the decoded file can peak above a master that looked clean.
- Set the final limiter in true-peak mode at -1 dBTP, or -2 dBTP for Spotify masters louder than -14 LUFS, and meter an encoded copy.
In this lesson (8)
Clean in the session, crackling on the phone
You bounce a master and the peak meter reads -0.1 . It sounds clean on your monitors. Weeks later, on a phone, the hi-hats fizz and the loudest snare hits carry a faint crackle. Nothing in your session changed.
Between your bounce and that phone, the file was turned back into a continuous wave and passed through a lossy codec. Both steps can create peaks your meter never showed you.
Why it matters
The meter reads samples, the speaker plays a wave
An ordinary peak meter reports the largest sample value. A digital file is a list of samples, but what reaches the speaker is the continuous wave those samples describe, rebuilt by the reconstruction filter in a converter. That wave can crest between two samples, higher than either of them. These are .
How far off a sample meter is depends on where the samples happen to fall. ITU-R BS.1770 gives the textbook case: a tone at a quarter of the , sampled 45 degrees away from its crests, reads 3 dB low. For short with a lot of high-frequency content, the standard says the under-read can commonly be several decibels.
A dense, heavily limited master is the worst case. Many samples sit right at the ceiling, so the rebuilt wave gets many chances to rise above it. Anything above full scale clips in whatever stage has to hold the signal within a fixed range: a digital-to-analog converter, a sample rate converter, or a decoder that writes fixed-point audio. EBU Tech 3343 lists exactly those places.
The science
True peak and what encoding adds
A meter estimates the continuous wave by . BS.1770 describes the reference method: raise a 48 kHz signal four times, to 192 kHz, with an interpolating , then take the largest absolute value. The result is reported in dBTP. Even this meter can miss a little, and the standard gives the worst case:
Here is the oversampling ratio and the highest frequency as a fraction of the sample rate. With and , that is about -0.69 dB. EBU R 128 therefore sets the production ceiling at -1 dBTP rather than 0: its guidelines explain that 1 dB of covers the possible under-read of a 4x meter.
Lossy encoding adds a second problem. An MP3, AAC or Ogg Vorbis encoder does not store your waveform. It stores an approximation that should sound the same, and the decoded wave is a different wave, which can peak higher than the original. EBU Tech 3343 names lossy coding as a cause of peaks above the sample level, and Apple's mastering guide warns that "levels that don't show overs on PCM can still cause when encoded" (Apple, 2021).
That is why ceilings for encoded delivery are lower. The EBU recommends -2 dBTP for the broadcast codecs MPEG-1 Layer 2 and Dolby AC-3. Spotify recommends below -1 dBTP in general and below -2 dBTP for masters louder than -14 , because louder tracks are more likely to distort when they are transcoded.
Try it
Meter the encoded file
0 of 6 steps done
On a loud master the encoded file often reads higher than the WAV, and on a dense chorus it can pass 0 dBTP when the WAV did not. With the lower true-peak ceiling, the encoded copy should stay below 0. On a Mac, Apple's AURoundTripAAC plugin compares your audio with its AAC encode and flags clipping in real time.
Common mistake
Trusting the sample meter
Setting the to 0 or -0.1 dB in sample-peak mode leaves no room for reconstruction or the codec. Your DAW still looks clean, because the overs appear later, in a converter or decoder you do not control.
The second mistake is expecting a lower ceiling to rescue an over-limited master. -1 dBTP gives the codec room. It does not restore the flattened transients. Spotify asks louder masters for the extra decibel because they are more prone to distortion when they are encoded.
Takeaway
Set the ceiling in true peak
Turn on true-peak detection in your final limiter and set the ceiling to -1 dBTP, or -2 dBTP for masters louder than -14 LUFS going to Spotify. Then check an encoded copy as well as the WAV, because the encoded file is the one your listener hears. The rest of the delivery chain is covered in the lesson on delivery specs.
dBFS
Decibels relative to full scale, the largest value a fixed-point file or converter can hold, so levels below it are negative numbers. Inside a floating-point mix a signal can read above 0 dBFS and stay intact; it clips where it leaves float or passes through a plugin that clips. dBTP is the same scale measured as true peak, and it can read above 0.
Inter-sample peak
A peak in the reconstructed waveform that falls between two samples, so the highest sample value under-reads it. True-peak meters find it by oversampling.
Sample rate
How many times per second a digital system measures the waveform. 48 kHz means 48,000 measurements a second.
Transient
The short, sharp start of a sound, like the crack of a snare or a pick hitting a string. It carries a lot of how hard a sound feels and how quickly the ear recognizes it.
True peak
The highest level a waveform reaches between samples once it is converted to analog or decoded from a lossy file. It can sit above the highest sample value, so meters estimate it by oversampling.
Oversampling
Running a process at a multiple of the session sample rate, then filtering and converting back. It gives distortion and clipping room to create high harmonics without folding them back as aliasing.
Low-pass filter
A filter that removes energy above a cutoff frequency and lets everything below it through.
Headroom
The space between the loudest peak in a signal and the level where it would clip.
Clipping
Cutting off the top of a waveform at a fixed level. Hard clipping flattens the peaks; soft clipping rounds them, which adds fewer harsh harmonics.
LUFS
Loudness Units relative to Full Scale. A loudness measurement weighted toward how the ear hears. Streaming services use integrated LUFS, an average over the whole track, to match playback level.
Limiter
A very fast compressor with a very high ratio. It stops peaks from going over a set ceiling.
3 questions
Check yourself
Sources
Show all 5Hide sources
- Apple. (2021). Apple Digital Masters [Technology brief]. www.apple.com/apple-music/apple-digital-masters/docs/apple-digital-masters.pdf
- European Broadcasting Union. (2023). EBU R 128: Loudness normalisation and permitted maximum level of audio signals. EBU. tech.ebu.ch/publications/r128/
- European Broadcasting Union. (2023). Tech 3343: Guidelines for production of programmes in accordance with EBU R 128. EBU. tech.ebu.ch/docs/tech/tech3343.pdf
- International Telecommunication Union. (2023). Recommendation ITU-R BS.1770-5: Algorithms to measure audio programme loudness and true-peak audio level. ITU. www.itu.int/rec/R-REC-BS.1770/
- Spotify for Artists. Loudness normalization on Spotify. support.spotify.com/artists/article/loudness-normalization/
Next lesson · 41 of 44
Mastering should translate the songMastering works on the whole stereo file, so it cannot move one instrument. What it can do is make the song hold together on every system it meets.Previous: Clipping can work. It can also cost you

Written by Virzy Guns
Songwriter and producer, founder of Virzy Guns Production. Now building HealingWave.