Mastering Metering Guide: LUFS RMS and True Peak
LUFS tells you how loud the master is perceived over time, RMS tells you how dense or energetic the audio feels, and true peak tells you whether inter-sample peaks may clip after conversion or streaming encoding. Use all three together: LUFS for playback loudness, RMS for density and crest-factor clues, and true peak for delivery headroom.
Mastering meters are useful, but they are easy to misuse. A loudness meter cannot tell you whether the chorus feels better. An RMS number cannot tell you whether the mix is emotionally finished. A true-peak reading cannot tell you whether the vocal is too bright. Meters show evidence. The mastering decision still comes from listening.
The fastest way to ruin a self-master is to chase one number without asking what the meter is actually measuring. A song can hit a common LUFS target and still sound weak. A song can have safe true peaks and still distort because the mix is harsh or over-limited. A song can show strong RMS and still feel small because the low end is eating the headroom. The meter should guide the diagnosis, not replace the diagnosis.
Want a release-ready master checked for loudness, translation, and headroom?
Book Mastering ServicesThe Short Version
LUFS is the best meter for perceived loudness because it is designed around how loud audio feels to listeners. Integrated LUFS measures the whole song or selected program over time. Short-term LUFS shows what is happening over a smaller window. Momentary LUFS reacts faster. Loudness range shows how much the program moves dynamically.
RMS is older and less playback-platform-focused, but it is still useful. It measures average signal energy. In mastering, RMS can help you understand density, body, and how far the peaks are above the average level. It is not weighted like LUFS, so it does not match perceived loudness as well, but it can still reveal whether a master is getting crushed.
True peak estimates peaks that may happen between digital samples. That matters because lossy encoding, sample-rate conversion, and playback reconstruction can create peak levels that are higher than the sample peaks you see on a normal digital peak meter. A master that shows 0 dBFS sample peak can still clip in real playback. True peak gives you a safer delivery check.
| Meter | What it tells you | Best use | Common mistake |
|---|---|---|---|
| Integrated LUFS | Perceived loudness across the whole song | Checking final playback loudness and comparing references at matched level. | Turning every song into the same number regardless of genre and arrangement. |
| Short-term LUFS | Loudness over a moving window | Seeing how verses, hooks, drops, and bridges compare. | Fixating on short-term jumps that are musically intentional. |
| RMS | Average energy and density | Understanding punch, body, and how hard the limiter is working. | Using RMS as a streaming target. |
| True peak | Estimated inter-sample peak level | Leaving enough headroom for encoding and playback reconstruction. | Assuming sample peak and true peak are the same thing. |
What LUFS Actually Measures
LUFS stands for Loudness Units relative to Full Scale. It is a loudness measurement used because normal peak meters do not describe how loud audio feels. A snare transient can peak high without making the whole song feel loud. A dense synth pad can feel loud without peaking as high. LUFS helps measure that perceived loudness more directly.
The modern loudness world is built around standards such as ITU-R BS.1770, which defines algorithms for measuring audio program loudness and true-peak audio level. Broadcast recommendations such as EBU R 128 use loudness normalization concepts for program exchange. Streaming platforms use their own playback behavior and normalization systems, but the central idea is the same: peak level alone is not enough to describe loudness.
Integrated LUFS is the number most people talk about because it describes the whole song or program. If you measure only one chorus, you are not reading integrated loudness for the entire track. If you include long silence at the start or end, you can skew the result depending on your meter and gating behavior. Measure the actual exported song area you intend to release.
Short-term LUFS is useful while mastering because it shows section balance. If the verse is much softer than the hook, that may be intentional. If the bridge drops too far and disappears on quiet playback, the short-term meter can confirm what you heard. Momentary LUFS reacts even faster and can show sudden loudness spikes, but it is not the number you should obsess over for final delivery.
Loudness range helps you understand macro-dynamics. A ballad, acoustic record, cinematic piece, or dynamic rock song may naturally have more range than a dense trap single. Less range is not automatically worse, and more range is not automatically better. The question is whether the dynamic movement supports the record and translates to the listener's playback environment.
What RMS Still Tells You
RMS stands for root mean square, and in audio it is used as a way to describe average signal energy. Before LUFS became the more common loudness conversation, many engineers used RMS readings to estimate how loud or dense a master was. RMS does not model human loudness perception as well as LUFS, but it is still useful as a second clue.
Think of RMS as density. If peak level stays similar but RMS rises, the audio is getting more constant. That can mean better sustain, more body, more saturation, more compression, or too much limiting. A dense pop master may show a higher RMS than a sparse acoustic song. That does not mean the acoustic song is wrong. It means the arrangements are different.
RMS is especially helpful when comparing similar versions of the same song. If a limiter push raises RMS but makes the drums smaller, you know the master is trading punch for density. If a low-end EQ move raises RMS but makes the song feel less loud, the low end may be taking headroom without improving perceived level. If saturation raises RMS slightly while the vocal becomes easier to hear, that may be a useful trade.
Do not use RMS as a platform target. Streaming platforms that publish loudness behavior talk in LUFS, not RMS. RMS is a diagnostic tool. It tells you something about energy, compression, and crest factor. It does not tell you whether the master will be normalized up or down on a given service.
What True Peak Actually Protects
A digital sample-peak meter shows the highest sample value in the file. True peak estimates what may happen between samples when the audio is reconstructed, converted, or encoded. This matters because the reconstructed waveform can exceed the sample values, especially when audio is limited close to 0 dBFS.
That is why true peak is usually expressed as dBTP. A true-peak limiter or true-peak meter oversamples or estimates the waveform between samples to catch peaks a normal sample meter can miss. This does not magically make a bad master good. It simply reduces the risk that the file clips after conversion or lossy encoding.
Spotify's public loudness normalization guidance is a useful example: it describes playback normalization around -14 dB LUFS for normal listening behavior and recommends keeping masters below -1 dB true peak, with a lower true-peak ceiling for louder masters to reduce extra distortion risk after encoding. That is platform-specific advice, not a universal law for every release, but it shows why true peak matters.
True peak should be checked at the end of the chain. If you put a true-peak meter before a limiter, it will not tell you what the final exported file does. If you sample-rate convert after limiting, recheck true peak after conversion. If you create MP3 or AAC preview files, check those too when distortion matters.
How to Meter a Master From Start to Finish
Metering should happen in a repeatable order. If you only look at the final number after ten processing moves, you will not know which move changed the loudness, density, or peak behavior. Check before and after the important stages so the meter teaches you what the processing is doing.
- Import a reference track and turn it down to match your mix by ear.
- Measure the unmastered mix for peak level, integrated LUFS, short-term loudness, RMS, and true peak.
- Fix mix issues before mastering if the vocal, low end, or harshness is already wrong.
- Add EQ or tonal shaping and watch whether loudness changed just because the tone changed.
- Add compression or saturation and compare RMS, LUFS, and punch at matched output level.
- Add limiting last and watch integrated LUFS, true peak, and whether transients collapse.
- Export the master, reimport it, and measure the printed file from start to finish.
- Check the printed file on multiple playback systems before deciding the numbers are finished.
- Create alternate versions only when the release, client, or platform need is real.
This order keeps the meter attached to decisions. You are not just asking whether the song hit a number. You are asking what changed, whether the change sounds better, and whether the final file is safe enough for release.
Common Meter Readings and What They Mean
| Reading | What it may mean | What to check next |
|---|---|---|
| Integrated LUFS is much quieter than references | The master may be too dynamic, too low in level, or not dense enough for the genre. | Match references by ear, then check whether the mix lacks midrange, saturation, or limiter level. |
| Integrated LUFS is loud but the song feels small | The limiter may be flattening drums, bass, or vocal depth. | Back off the limiter, fix low-end headroom, or use gentler staged processing. |
| RMS rises but perceived loudness barely improves | Low end may be eating headroom without adding audible loudness. | Check sub balance, bass sustain, kick peaks, and low-mid buildup. |
| True peak is over the delivery ceiling | Inter-sample peaks may clip after conversion or encoding. | Lower limiter ceiling, use true-peak limiting, and recheck after export. |
| Short-term loudness jumps hard in the hook | The hook may be more dense, more limited, or intentionally explosive. | Decide whether the jump feels musical or whether the verse/hook balance needs automation. |
Why One LUFS Target Can Mislead You
It is tempting to ask, "What LUFS should my master be?" The honest answer is that it depends on the genre, arrangement, release format, reference tracks, and how much distortion or compression the song can handle. A quiet acoustic song, aggressive rap single, EDM drop, metal track, and piano ballad should not all be mastered with the same density just because one streaming number is popular.
Streaming normalization also happens during playback, not as a command that changes what you must master to. If a platform turns a loud master down, the loud master may still sound different from a more dynamic one because limiting, clipping, saturation, and tonal balance are printed into the file. Normalization changes playback gain. It does not restore lost transients.
The better question is: how loud can this song be before it gets worse? Raise level in small steps. Level-match your comparisons. If the chorus gets exciting without losing drums, vocal depth, low-end movement, or top-end smoothness, the master can handle the push. If every extra half dB makes the track flatter, stop.
The LUFS targets for every streaming platform guide is useful for platform behavior, but even target guides should not become a substitute for listening. Use them to avoid delivery mistakes, not to force every song into one loudness box.
How LUFS, RMS, and True Peak Work Together
These meters become more useful when you read them together. Integrated LUFS tells you the perceived loudness of the whole track. RMS gives a sense of average energy and density. True peak shows delivery headroom risk. If all three move in a healthy direction and the song sounds better, you are probably improving the master. If one number improves while the song gets worse, the meter is warning you about the tradeoff.
For example, imagine a master that moves from -12 LUFS to -9 LUFS. The RMS rises, but the true peak keeps hitting the ceiling and the kick loses punch. That is not automatically a win. The song is louder, but it may be less impactful. Another master might stay around -11 LUFS, keep safer true peak headroom, and feel larger because the transients survive. The less loud master can be the better record.
Now imagine a vocal-led pop song where integrated LUFS barely changes after a high-frequency EQ lift, but the lyric becomes easier to hear. That might be a good move because clarity improved without extra limiting. Or imagine a low-end boost that raises RMS but makes the song feel quieter because the limiter clamps down. That is a bad trade. The meter explains why.
Metering Your Limiter
The limiter is where many mastering numbers change quickly. It controls peaks, raises average level, and can reshape transients. Because it is powerful, it can also hide problems until the song is exported. Watch integrated LUFS, RMS, gain reduction, and true peak while adjusting the limiter, but listen harder than you look.
A good limiting move makes the song feel finished without obvious damage. A bad limiting move makes the snare smaller, smears the vocal, distorts the low end, or makes the top end spit. If the limiter has to work too hard, do not assume the limiter is bad. The mix may have too much sub, too much kick transient, harsh cymbals, or an uncontrolled vocal peak stealing headroom.
The mastering limiter settings guide covers ceiling, attack, release, gain reduction, and distortion checks in more detail. For metering, the key is to recheck the printed file. The number inside the plugin is not proof until the exported master measures correctly.
Metering Before You Send a Song for Mastering
If you are sending a mix to a mastering engineer, do not overprocess it just to make the rough bounce loud. A good mastering engineer needs a clean mix with enough headroom and no accidental limiter damage. You can send a loud reference bounce separately if you like the direction, but the main mix should not be clipped, normalized, or crushed just to impress someone.
Use meters to check that the mix is not accidentally peaking, clipping, or collapsing before delivery. Peak level does not need to hit a magic number, but the file should not be clipping and should have room for mastering decisions. If the mix bus chain is part of the sound, print a version with it and communicate that clearly. If it is only there for loudness, consider bypassing it for the mastering version.
The DIY mastering checklist is useful if you are doing the final stage yourself. If you are handing the song off, use metering to avoid preventable technical problems, then trust the engineer to make the final loudness calls.
Metering Mistakes That Cost Translation
The biggest mistake is measuring without level matching. Louder almost always sounds better for a moment. If you compare a processed master against an unprocessed mix and the processed version is louder, you may think the processing helped when it only raised gain. Match levels before judging tone, punch, width, and vocal clarity.
The second mistake is measuring only the loudest section. The chorus matters, but the whole song matters too. A song can have a strong hook and a weak verse, or a safe verse and a distorted final chorus. Measure the printed file from beginning to end and watch section behavior during playback.
The third mistake is ignoring the codec or sample-rate stage. If you export, convert, or create a lossy preview, peaks can change. Check the final delivery file or at least the exact master file you are submitting. If true peak rises after conversion, adjust the limiter ceiling or workflow and export again.
The fourth mistake is using loudness normalization tools as mastering. A normalizer can change level, but it does not fix harshness, low-end balance, stereo width, vocal placement, or distortion. The LUFS normalizer guide explains what should be fixed before a loudness tool touches the file.
A Simple Mastering Metering Checklist
- Measure the unmastered mix before processing so you know the starting point.
- Compare references at matched loudness, not raw volume.
- Watch integrated LUFS for whole-song loudness, not only chorus loudness.
- Use short-term LUFS to compare sections and transitions.
- Use RMS to understand density and limiter pressure.
- Use true peak to protect delivery headroom after limiting and export.
- Check the printed file, not only the live plugin chain.
- Listen on real playback systems before deciding the meter numbers are good.
- Stop pushing loudness when the song gets smaller, harsher, or less emotional.
Final Take
LUFS, RMS, and true peak are not competing meters. They answer different questions. LUFS asks how loud the song is likely to feel. RMS asks how dense the signal is. True peak asks whether the final file has enough technical headroom. A good master respects all three without worshiping any one of them.
Use the meters to catch problems your ears can miss, especially after long sessions. Then return to the record. Does the vocal still feel present? Does the hook still lift? Does the low end translate? Do the drums still hit? Does the master survive quiet playback? If the answer is yes and the technical readings are safe for the release path, the numbers have done their job.
FAQ
Is LUFS better than RMS for mastering?
LUFS is better for perceived loudness and platform-related loudness decisions. RMS is still useful for judging average energy and density, but it should not replace LUFS for modern mastering checks.
What true peak ceiling should I use?
A common safe starting point for streaming is around -1 dBTP, with more room for very loud masters. Always check the release path and remeasure the exported file after any conversion.
Should every master be -14 LUFS?
No. -14 LUFS is associated with some playback normalization behavior, but it is not a universal mastering command. Master the song to sound right, then make sure the delivery file is technically safe.
Why does my master sound quieter even when the LUFS number is loud?
The master may have too much low-end energy, crushed transients, harsh top end, or poor tonal balance. LUFS measures loudness, but tonal balance and punch still affect perceived impact.
Do I need a paid loudness meter?
You need a reliable LUFS and true-peak meter, but it does not have to be expensive. The important part is understanding integrated loudness, short-term loudness, RMS, and true peak in context.
Should I meter before or after exporting?
Do both. Meter during mastering to make decisions, then reimport and meter the printed export so you know the final file has the loudness and true-peak behavior you intended.
