Tutorials

What Is Audio Clipping and How to Prevent It? Gain Staging, Headroom and Safe Levels

Why a recording crackles when it gets too loud, what 0 dBFS means, and how to set your levels so it never happens.

By Yunus S. Published Updated

Audio clipping happens when a signal tries to go above 0 dBFS, the highest level a digital system can store. The tops of the waveform are cut off, which sounds like crackling or harsh distortion, and the lost detail cannot be restored. You prevent it by recording with headroom: set the input gain so your loudest moments peak between -12 and -6 dBFS.

In Short

The short version:

  • Clipping means the waveform is cut flat at 0 dBFS; it sounds crackly and harsh.
  • It cannot be fully repaired afterwards, so prevention matters.
  • Record with peaks between -12 and -6 dBFS; in 24-bit there is no penalty for recording lower.
  • When you edit, check the peaks every time you add gain, boost with EQ or mix in a second track.
  • Keep the true peak of the finished file at or below -1 dBTP.

What Is Audio Clipping?

Digital audio stores sound as numbers, and those numbers have a fixed maximum. That maximum is called 0 dBFS (decibels relative to full scale); every other level is measured downwards from it as a negative value, such as -6 dBFS or -20 dBFS. Stanford’s CCRMA gives a short definition of the dBFS scale .

When a signal is pushed past 0 dBFS, there are no larger numbers left. Every sample that should be higher is stored at the maximum instead, so the rounded top of the wave becomes a flat line. That sudden corner adds harsh, high-pitched distortion that was not in the original sound. On speech it is heard as crackling on loud syllables; on music, as a gritty, crushed quality.

Digital audio meter showing red clipping during a recording
When the meter reaches the top and turns red, the peaks of the waveform are being cut off.

Why Digital Clipping Sounds Worse Than Analog

Tape machines and valve amplifiers also distort when they are overloaded, but they do it gradually: the wave is rounded off, which adds a warmth some engineers use on purpose. Digital systems have no such soft zone. Below 0 dBFS the signal is perfectly clean; above it, the distortion is immediate. Some producers clip drums or bass deliberately for an aggressive sound, but that is a controlled choice made with tools designed for it, not something that should happen to a voice recording.

Where Clipping Happens

Clipping can enter a recording at four different points, and the fix depends on which one it is:

  • At the microphone preamp or interface. The input gain is too high, so the signal is already clipped before it reaches the computer. Nothing you do afterwards can fix it.
  • While editing or mixing. Adding gain, boosting with EQ or mixing two tracks together raises the peaks. A file that was safe can go above 0 dBFS after a single change.
  • On export. A mix that goes above 0 dBFS inside the editor is cut off when it is saved as a 16-bit or 24-bit file.
  • After conversion to MP3 or AAC. Lossy encoding changes the waveform slightly, and peaks that were just under 0 dBFS can end up above it. This is why finished files should leave about 1 dB of room.

Gain Staging: Setting the Right Recording Level

Gain staging means choosing a sensible level at every point in the chain, starting with the input gain on your interface or USB microphone. The aim is simple: the loudest thing you will record should stay well below 0 dBFS.

  1. Speak, sing or play the loudest part of what you are about to record.
  2. Turn the input gain until those moments peak between -12 and -6 dBFS on the meter.
  3. Leave it there. Normal passages will sit around -18 dBFS, which is fine.

The space between your peaks and 0 dBFS is called headroom. In 24-bit recording the noise floor is so low that a quiet-looking recording loses nothing: you can raise it later without a penalty. A recording that clipped, on the other hand, stays damaged. If your interface has only a clip light, turn the gain down until it never flashes on your loudest sounds, then a little further.

Hand adjusting the gain control on a studio audio interface
Clipping is prevented at the first control in the chain: the input gain.

How to Avoid Clipping When You Edit and Mix

A clean recording can still be clipped later. These four steps keep the peaks under control while you work.

Step 1: Measure First

Open the file in the LUFS & Peak Analyzer. It shows the sample peak and the true peak; if either is at or above 0 dB, the file is already clipped or will clip when it is converted.

Step 2: Cut Before You Boost

Rumble below 80 to 100 Hz uses up headroom without adding anything you want to hear. Removing it with the Low Cut filter in the Sound Optimizer often lowers the peaks by a decibel or two. When you use EQ, turning down what is too much is safer than boosting what seems to be missing.

Step 3: Control the Loudest Moments

A compressor turns down the highest peaks so that the whole recording can be raised safely. In the Sound Optimizer, set the Compression Ratio to about 3:1 and lower the Compressor Threshold until only the loudest moments are reduced.

Step 4: Raise the Level and Check Again

Raise Volume / Gain while watching the Peak meter and keep it at or below -1 dB. If the export would still go above 0 dBFS, the tool lowers the whole file to -1 dBFS instead of letting it clip. Measure the exported file once more.

To choose a loudness target and work out how much gain is safe, see How to Normalize Audio Without Clipping.

Can Clipped Audio Be Fixed?

Only partly. When the top of a wave is cut off, the missing shape is not stored anywhere, so no tool can know exactly what it was. Dedicated de-clipping software estimates the lost peaks from the audio around them, and on light clipping the result can be acceptable. On heavy clipping it cannot remove the distortion. Turning the file down afterwards makes it quieter, not cleaner. If you can record again at a lower level, that is always the better fix.

Signs of Clipping and What to Do

What you notice Likely cause What to do
Crackling only on loud words or hits Input gain too high Record again with peaks at -12 to -6 dBFS
Flat tops visible in the waveform Recording or export above 0 dBFS Lower the level before exporting; record again if it is in the source
Clean WAV, distorted MP3 Peaks too close to 0 dBFS before encoding Lower the WAV by 1 dB and encode again
Distortion after mixing voice and music Combined level too high Lower the music or the overall level

Check Your Peaks

Open a file in the LUFS & Peak Analyzer and see its sample peak and true peak in a few seconds.

Open the LUFS & Peak Analyzer

Free, no sign-up, and your files never leave your device.

Frequently Asked Questions

What does audio clipping sound like?

Clipping sounds like crackling, buzzing or a harsh, crushed tone on the loudest parts of a recording, such as shouted words, laughs or drum hits. Quiet passages usually sound normal.

What causes audio clipping?

A signal that is too loud for the system that carries it. In digital audio this means trying to go above 0 dBFS, usually because the input gain was set too high during recording or too much gain was added while editing.

Can crackling in an audio recording be fixed later?

Light clipping can sometimes be improved with dedicated de-clipping software, which estimates the missing peaks. Heavy clipping cannot be repaired, because the original shape of the waveform is lost. Recording again at a lower level is the reliable fix.

What level should I record at to avoid clipping?

Set the input gain so your loudest moments peak between -12 and -6 dBFS. That leaves enough headroom for an unexpected laugh or a louder note, and in 24-bit recording there is no quality penalty for recording at this level.

What is headroom?

Headroom is the space between the highest peak of your audio and 0 dBFS. Recording with 6 to 12 dB of headroom means sudden loud moments have somewhere to go without clipping.

What is gain staging?

Gain staging is setting a sensible level at each stage of the signal chain, from the microphone preamp to the final export, so that nothing overloads and nothing is so quiet that noise becomes a problem.

Does a limiter prevent clipping?

Yes. A limiter is a very fast compressor with a fixed ceiling, for example -1 dB, that turns down any peak before it can pass that level. Used lightly it is hard to hear; if it has to reduce the level a lot, the sound becomes flat and distorted in its own way.

Is clipping ever used on purpose?

Yes. In some styles of music, producers clip drums or bass deliberately with soft-clipping tools to make them sound louder and more aggressive. That is a controlled effect, different from accidental clipping in a recording.