Updated October 2026 · Levels and metering explained
A VU meter shows an average level, reacting slowly to match how loud a sound seems, while a peak meter shows the single loudest moment, reacting quickly. Watch the peak meter to avoid clipping, and use the VU-style average to judge how loud a passage feels. Use both together, because each answers a different question.
Level meters come in several designs, and two of the most common are the VU meter and the peak meter. They look similar, but they show different things. A VU meter shows an average level, while a peak meter shows the loudest moment. Beginners who watch the wrong one can set levels that clip or that sound too quiet, so it is worth understanding what each meter measures.
This post explains how each meter works, why they disagree, which one to watch for each job and how to use them together.
A VU meter is designed to show something close to the loudness a listener perceives. It reacts slowly, averaging the level over a fraction of a second, so the needle does not jump with every spike. That makes it useful for judging how loud a passage feels, but it can hide short, sharp peaks that will clip.
Many digital meters offer a VU-style display alongside a peak display, and the VU-style reading is often shown as an average level.
A peak meter reacts quickly to the loudest moment in the signal. It shows the highest level reached, and it is the meter to watch for clipping. A peak meter can look busy on speech, because the peaks appear and disappear with every word, but the reading is essential for avoiding distortion.
A speech signal has short, loud peaks and a lower average. The peak meter shows the peaks, which can be well above the average, and the VU meter shows the average, which sits lower. Both readings are correct, and they answer different questions. A speech recording can show peaks near the top on the peak meter while the average sits comfortably in the middle.
During a recording session, the two meters can tell you different and useful things. The peak meter warns you when a loud word is about to clip. The average meter tells you whether the performance feels consistent from one take to the next. Watching both as you record lets you keep the peaks safe and the feel steady, without chasing every change in the peak meter.
A practical habit is to run the first few takes with both meters in view, noting where the average sits for a verse and a chorus. Later takes can then be compared with those reference points, which makes it easier to notice when a singer or speaker has drifted.
In editing, the average meter helps you match the level of one section to another, while the peak meter confirms that no edit has pushed a moment into clipping. A section that looks correct on the average can still hide a loud spike, and a section that looks quiet on the peaks may feel too low in the average. Use both, and listen to the join between sections before making a final decision.
Make a note of the average level of a section that sounds right, and use it as a target for the rest of the edit. A reference point is more useful than an abstract number. Write the reference beside the section it came from, so you can see the context when you return to the edit.
Meters differ widely in how they calculate and display level, and the names are used inconsistently across products. Check what each meter in your software measures before comparing readings, and do not assume that two meters labelled the same will agree.
The finished file is where level matters most. The LUFS meter shows the loudness and true peak of an export, and the Vocal Checker shows clipping and level problems in a recording you own. For the decibel scale behind all these readings, read the decibels post in this series.
Integrated LUFS, true peak and streaming targets. Runs in your browser, nothing is uploaded.
The peak meter, because it shows the loudest moment.
Speech has short, loud peaks and a quieter average. Both readings are correct.
Not on its own. Use a peak reading to check for clipping.
Ideally yes, since each answers a different question about the level.
