Skip to content

PLR and PSR: peak-to-loudness ratios explained

Send feedback

Last updated 7 min readTags:MasteringLoudnessDynamics

Learn how PLR and PSR compare peaks with loudness, why whole-track readings can miss dense sections, and how to check whether limiting has reduced punch.

Peak-to-loudness ratio (PLR) is the difference between the highest true peak and the integrated loudness of a track. Peak-to-short-term loudness ratio (PSR) makes a similar comparison over a moving three-second window.

PLR gives you one figure for the track; PSR shows how the relationship between peaks and loudness changes during it. Both can be useful when you check compression or limiting, but neither measures sound quality or tells you how much processing was used.

How to calculate PLR and PSR

Measurement Calculation Period measured
PLR Maximum true peak minus integrated loudness The whole track, or another selected passage
PSR Maximum true peak within a three-second window minus short-term loudness for that window A moving three-second window

The EBU defines PLR using programme loudness and maximum true peak. For PSR, Sonible documents the calculation using the peak and loudness within the same three-second window. Check your meter's documentation when you compare results.

For example, a track with a maximum true peak of −0.8 dBTP and integrated loudness of −9.0 LUFS has:

PLR = −0.8 − (−9.0) = 8.2 dB

If one three-second passage has a maximum true peak of −1.0 dBTP and short-term loudness of −7.0 LUFS, its PSR is 6.0 dB.

A lower number means less difference between the peak and the measured loudness. A higher number means more difference. Neither is automatically preferable: a sustained organ part and a sparse drum recording have quite different peak patterns before any processing.

Why PLR can miss a dense chorus

PLR combines the highest peak with an integrated loudness measurement. It does not tell you which sections have the smallest gap between peaks and loudness.

Two tracks can therefore have the same PLR and quite different choruses. Here is an illustrative example, with invented figures:

Measurement Track A Track B
Whole-track maximum true peak −1 dBTP −1 dBTP
Integrated loudness −12 LUFS −12 LUFS
Whole-track PLR 11 dB 11 dB
True peak in the selected chorus window −1 dBTP −1 dBTP
Short-term loudness in that window −10 LUFS −7 LUFS
PSR in that window 9 dB 6 dB

Track B's selected chorus window has a smaller peak-to-loudness gap, despite the identical PLR. That is a useful place to compare the sound before and after limiting. The figures alone do not establish why the gap is smaller.

Quieter sections can affect integrated loudness, but the calculation is not a simple average of everything in the file. Integrated loudness uses gating, which excludes sufficiently quiet material. An extended silence does not simply pull the reading down.

PSR and “peak above the loudest three seconds”

Check which peak a meter uses. These calculations can produce different results:

  • Local PSR: the peak within a three-second window minus the loudness of that same window.
  • Whole-track peak minus maximum short-term loudness: the highest peak anywhere in the track minus the loudness of its loudest three-second window.

Suppose the track's highest peak is −0.5 dBTP, but the peak within its loudest three seconds is −2.0 dBTP. If that window measures −7.0 LUFS, the whole-track calculation gives 6.5 dB; local PSR gives 5.0 dB.

The first figure uses a peak from elsewhere and makes the local gap appear larger. It should not be treated as interchangeable with the minimum PSR from a moving-window meter. Nor must the lowest PSR occur in the loudest window: a quieter, sustained passage may have an even smaller local gap.

If a report labels a value “peak above the loudest 3 s”, read its calculation before comparing it with PSR targets from another tool.

What does a low value mean?

A low PLR or PSR means the measured peak is relatively close to the measured loudness. Heavy compression, limiting or clipping can cause this, but so can the source material and arrangement.

The Dynameter guide recommends checking persistent PSR readings below 8 dB for excessive processing. It also identifies sustained sounds, such as pads and held notes, as possible sources of naturally low readings. Its 8 dB recommendation is a listening guideline, rather than a universal pass mark.

The useful question is whether processing has made the track sound worse. A low reading alongside dull drum attacks, audible pumping or roughness warrants a comparison. The same reading in a clean, sustained passage may require no change.

A high reading does not guarantee a good mix either. One unusually large hit can raise PLR without improving the rest of the track. Do not add sharper transients simply to increase the number.

How to check whether limiting has reduced punch

Start with a chorus, drop or other passage where the drums seem weaker than you intended. If you have a PSR history, inspect that passage and any other sections with persistently low readings.

  1. Reduce the drive into the final limiter by about 2 dB. Leave its output ceiling unchanged. If your limiter uses a threshold to set the amount of limiting, raise that threshold instead.
  2. Match the listening volume. Turn down the louder version with a gain control after its processing, or use a loudness-matched comparison tool. Do not feed the quieter version harder into the limiter to match it.
  3. Compare the same passage. Listen for a clearer initial kick or snare hit, less cymbal roughness, and less unwanted movement in the backing when the kick lands.
  4. Keep the change if it sounds better. The higher reading is only supporting information. If the less limited version loses the intended density without a useful improvement, reconsider the change.

This is a starting test, rather than a prescribed correction. You may find that 1 dB is enough, or that the audible problem comes from an earlier processor.

If there is no final limiter, inspect any clipping, saturation or bus compression you have applied. Compare those stages one at a time at matched volume. A final output trim cannot restore attacks already removed earlier in the chain.

PLR, PSR, crest factor and loudness range

These measurements answer different questions:

Measurement What it compares Useful for
PLR Maximum true peak and integrated loudness A summary of the peak-to-loudness gap over the measured audio
PSR Local true peak and three-second loudness Changes in the peak-to-loudness gap during playback
Crest factor Peak level and RMS level, in a typical audio meter The peak-to-average relationship over the meter's chosen interval
Loudness range (LRA) The statistical spread of short-term loudness Broader changes between quieter and louder passages

MeterPlugs explains why RMS-based crest factor and loudness-based ratios give different results. Do not transfer a preferred crest-factor value directly to PLR or PSR.

A song can have a quiet verse and a loud chorus, giving it substantial loudness variation, while the chorus itself has a small peak-to-loudness gap.

Does turning the track down change PLR?

An ordinary, constant gain adjustment moves peak level and loudness by the same amount. It therefore leaves PLR and PSR unchanged, provided it introduces no clipping and does not change which audio passes the integrated loudness gate.

For example, reducing the earlier track by 3 dB changes its peak from −0.8 to −3.8 dBTP and its integrated loudness from −9.0 to −12.0 LUFS. Its PLR remains 8.2 dB.

The same principle applies when playback normalisation uses a constant gain change. Turning down a heavily limited track does not restore the original drum attacks. Streaming playback rules also depend on the service and setting; Spotify, for example, documents both gain adjustments and a limiter for its Loud setting.

Is there a good PLR for every mix?

There is no single value that establishes whether a mix is ready to release. Use references with a similar arrangement and intended sound, measure them with the same tool, and compare them at matched loudness.

If your track has a lower PLR, check whether that difference is audible and undesirable. If it sounds right, there is no reason to revise it merely to match the reference's number.

Sources

    Type to search titles, headings and summaries.