Peak vs RMS: Differences and How to Use Each Meter
Last updated 7 min readTags:MasteringLoudnessDynamics
Peak measures the highest signal level; RMS measures level over time. Learn how to compare them, calculate crest factor and check changes to your mix.
Peak level measures the highest signal amplitude. RMS measures the signal's effective level over a period of time. Peak is useful for checking headroom; RMS helps you compare the average level of similar sounds or passages.
Two mixes can both peak at −1 dBFS and have quite different RMS readings. One might reach that peak on an occasional snare hit, while the other maintains a high level throughout. The peak reading alone cannot tell you which will sound louder.
| Measurement | What it measures | Main use |
|---|---|---|
| Sample peak, in dBFS | Highest absolute sample value | Check sample headroom and possible overloads |
| True peak, in dBTP | Estimated highest level of the reconstructed waveform | Check peaks that sample metering can miss |
| RMS, usually displayed in dBFS | Effective signal level over a chosen time window | Compare average level using consistent meter settings |
| Crest factor, in dB | Ratio of peak amplitude to RMS, expressed in decibels | Assess how far peaks rise above the average |
| LUFS | Loudness estimate with frequency and channel weighting | Compare loudness and check delivery requirements |
What does a peak meter measure?
A sample-peak meter checks the audio samples and reports the largest absolute value. Its scale is relative to digital full scale: a peak of −6 dBFS is 6 dB below the full-scale sample reference.
This makes it useful for checking levels through a recording or mixing chain. A brief drum hit may produce a high peak even when the rest of the passage is relatively quiet.
Peak metering has two limits worth remembering:
- A low current peak does not rule out earlier clipping. Distortion from an overloaded recording or processor can remain after you turn the signal down.
- Sample peaks can miss higher levels between samples. A true-peak meter estimates these peaks in the reconstructed waveform. The distinction is part of ITU-R BS.1770.
For final exports, check the rendered file with a true-peak meter, particularly if you have worked close to full scale.
What does RMS measure?
RMS stands for root mean square. The calculation squares the sample values, averages them over a time window, then takes the square root. This avoids the positive and negative parts of a waveform cancelling each other out in a simple average.
RMS is an amplitude measure related to average signal power. In production, it is useful for seeing how consistently a sound maintains its level.
A short click and a sustained synth note might have the same peak. Over a window that includes the click and its decay, the click will usually have a lower RMS because much of that window contains quieter audio. The sustained note maintains a higher level for longer.
The time window changes the result
An RMS reading over a few hundred milliseconds describes a short passage. An RMS calculation across an entire track also includes its quieter sections and any silence within the measurement.
That matters when you compare tracks. A long, quiet introduction can lower a whole-track RMS reading even if the chorus is heavily limited. Compare equivalent passages, such as one chorus against another, with the same window and channel settings.
Unweighted RMS also treats frequencies equally. It does not account for differences in hearing sensitivity, so equal RMS readings do not guarantee equal perceived loudness. Nor does a high RMS reading on a bass track, by itself, establish that it is masking a vocal.
What does the difference between peak and RMS mean?
The ratio of peak amplitude to RMS is called crest factor. With both levels expressed in decibels against the same amplitude reference:
Crest factor = peak level − RMS level
Here is an illustrative comparison. Both passages use the same measurement duration and channel configuration, with RMS referenced to full-scale sample amplitude.
| Passage | Sample peak | RMS | Crest factor |
|---|---|---|---|
| A | −1 dBFS | −13 dBFS | 12 dB |
| B | −1 dBFS | −9 dBFS | 8 dB |
Passage B has a higher average level relative to its peak. These readings do not reveal how it was produced. Heavier limiting is one possibility; sustained instruments, a denser arrangement or a different tonal balance could also contribute.
Crest factor is most useful when you compare the same passage before and after processing. If the ratio falls, the peaks have become smaller relative to the RMS level. Whether that sounds better depends on the material and the intended result.
It is not a measurement of how much musical dynamics remain. A sustained tone can have a low crest factor without any compression. A track can also have a high whole-track crest factor because of one unusually large transient or a long quiet section. See crest factor for more on interpreting the measurement.
Check the RMS calibration before subtracting
Some RMS meters apply an offset so that a full-scale sine wave reads 0 dBFS. Without that offset, its mathematical RMS is approximately −3.01 dBFS. RME's DIGICheck documentation, for example, specifies a full-scale sine as its RMS reference.
On a sine-calibrated display, simply subtracting RMS from sample peak gives a result about 3 dB below the mathematical crest factor. Check the meter's documentation, or use its dedicated crest-factor reading if the calculation is documented.
For reliable comparisons, keep the meter, calibration, time window and channel treatment consistent. A mono downmix can produce different results from separate left and right measurements.
How do gain, compression and limiting affect the readings?
A gain change moves peak and RMS together. If you turn a signal down by 3 dB without other processing, both readings fall by 3 dB. Crest factor stays the same.
Compression can change their relationship. A fast attack may reduce transients relative to the rest of a sound. A slower attack may preserve the initial hit while reducing what follows. The result depends on the source and settings; compression does not always reduce crest factor. FabFilter's guide to compressor controls includes audible examples of these changes.
Limiting can raise average level while holding peaks near a ceiling. Once a limiter is reducing peaks, more input gain can increase RMS without a corresponding rise in output peak. The exact RMS increase depends on the audio and the limiter. An extra 3 dB of input does not guarantee an extra 3 dB of RMS.
Listen for weaker drum attacks, pumping or distortion as you increase the level. A smaller crest factor is not automatically a fault, but it should accompany a result you prefer when compared at matched loudness.
Is RMS the same as LUFS?
No. Ordinary unweighted RMS describes signal level without frequency weighting. LUFS uses K-weighting and combines channel energy to estimate perceived loudness more closely.
Integrated LUFS measures a programme or selected passage with gating that excludes sufficiently quiet blocks. In EBU Mode, momentary and short-term LUFS use 400 ms and 3-second windows respectively, without that gating. These distinctions are specified in EBU Tech 3341.
Use LUFS for loudness comparisons and delivery checks. It remains an estimate of perception; two tracks with the same LUFS reading may still feel different in loudness.
How to use peak and RMS when comparing mixes
- Choose equivalent passages. Compare a representative chorus with a similar section of your reference. Avoid comparing a sparse introduction with a full arrangement.
- Use consistent measurement settings. Reset held readings and check the RMS calibration, time window and channel configuration.
- Compare loudness. Use LUFS to establish the level difference, then adjust playback gain for the listening comparison.
- Check the peak-to-RMS relationship. A higher RMS at the same peak means a smaller crest factor. It does not establish whether the cause is processing, instrumentation or arrangement.
- Test the relevant change. If one snare hit sets the maximum peak, adjust that hit and listen again. If limiting weakens the drums, reduce the drive and compare at matched loudness. If excess bass seems responsible, check the tonal balance before applying more limiting.
A mix that sounds quieter than its reference does not necessarily need a higher ceiling. Peak and RMS can show where the level difference lies; a matched-loudness comparison tells you whether the next adjustment improves the sound.