LUFS vs true peak explained in practical terms: LUFS describes perceived loudness over time, while true peak estimates the highest level the reconstructed audio waveform can reach between digital samples. They answer different questions. A loudness reading cannot tell you whether a transient will overload a playback or encoding stage, and a peak reading cannot tell you how loud the full track sounds. Check both when you are evaluating a mix or master, and do not treat either number as a universal target.
LUFS means Loudness Units relative to Full Scale. It is a unit used for loudness measurement, not a peak ceiling. Modern loudness meters following ITU-R BS.1770 apply frequency weighting and channel weighting, then calculate loudness over a selected interval. The integrated reading summarizes the measured programme; short-term and momentary readings describe shorter windows. The exact meter and measurement mode matter, so compare readings made with compatible standards and settings.
Integrated LUFS is useful for comparing whole tracks or checking a delivery specification. It does not show how high the loudest individual waveform peak gets. Two tracks can have the same integrated LUFS and very different peak levels because their arrangements, transient shapes, compression, and dynamic range differ. A dense, heavily limited master and a more transient master can produce similar average loudness while reaching different maxima.
A loudness number is also not a quality score. Raising integrated loudness with limiting or compression can reduce crest factor and change transients; turning a track down does not automatically make its dynamics better or worse. Make musical decisions by listening, and use the meter to describe the signal or meet a stated delivery requirement. The MixDiagnose LUFS checker can report loudness measurements for a rendered file, but it cannot decide whether the balance or dynamics sound right.
Digital audio stores samples at discrete points in time. A digital-to-analog reconstruction filter produces a continuous waveform from those samples. That reconstructed waveform can reach a maximum between sample points that is higher than the largest stored sample value. A sample-peak meter reports the highest sample; a true-peak meter estimates the inter-sample maximum, commonly expressed in dBTP (decibels relative to digital full scale).
True-peak measurement is defined in ITU-R BS.1770. In practice, meters use oversampling and filtering to estimate the reconstructed peak. It is an estimate, not a direct measurement of every playback device's analog output. The result depends on the meter's implementation and settings, so use a standards-conforming true-peak meter for delivery checks rather than inferring true peak from a sample-peak display.
True peak matters when a signal is converted, encoded, processed, or played back. Lossy encoding can change waveform peaks, and a downstream stage with insufficient headroom can clip. A low sample peak does not prove that true peak is equally low. Conversely, a true-peak number alone says nothing about the programme's average loudness or whether it is musically balanced.
Consider two renders with the same integrated LUFS. One has sharp drum and percussion transients; the other has been limited more heavily. Their integrated loudness may be alike, while the first may have higher peaks or the second may have less peak-to-average difference. LUFS does not identify the maximum waveform peak. True peak does not describe the loudness through the track.
The reverse comparison is just as important. Two files can share the same true-peak ceiling but have very different integrated loudness. A sparse acoustic recording and a dense electronic arrangement may both peak at the same dBTP value without sounding similarly loud. Lowering the master output changes both peak and loudness readings by the same gain amount, but it does not alter the relationship between their dynamics.
Do not use a single fixed LUFS value as a mixing target unless a client, broadcaster, platform, or delivery brief specifies it. Streaming services may apply loudness normalization, and their behavior and settings can vary. Normalization is not a reason to crush a mix to a particular integrated value. Preserve the dynamics and transient shape that serve the music, then meet the actual delivery requirements.
The MixDiagnose spectral analyzer provides a separate view of frequency energy. Spectrum, loudness, and peak measurements describe different properties; none replaces listening or the others. Use each result to narrow a question, then verify the musical result in context and on more than one playback system when practical.
Reading dBFS as LUFS. Sample peak in dBFS and programme loudness in LUFS are different measurements. A peak meter near full scale does not mean the whole track is loud, and an integrated LUFS value does not reveal the sample peak.
Assuming sample peak equals true peak. The reconstructed waveform may exceed its stored sample values. If true peak is part of the delivery requirement, use a meter that measures it.
Chasing a platform number without a brief. Loudness-normalization targets are not universal mastering instructions. Confirm the current delivery specification and prioritize the sound of the work rather than pushing every master toward one number.
Treating a meter as a verdict. Metering can reveal level and headroom issues, but it cannot diagnose every cause of harshness, masking, weak translation, or distortion. Pair the measurement with critical listening and, when relevant, a reference track at matched loudness.
LUFS tells you about perceived programme loudness over time; true peak estimates the highest level the reconstructed waveform may reach. Measure both when the task calls for both, but never swap one for the other or mistake a reading for a musical goal. Follow the actual delivery specification, listen for processing artifacts, and verify the final render.
Upload a mix to MixDiagnose to review loudness, peak, and other mix measurements before you make the next decision.