AUDIO QC BLOG / 8 MIN READ

How do I know if my audio is clipping?

Clipping is the crackly, fuzzy distortion you get when your sound is too loud for the system to hold. Here is how to spot it by eye, by ear and on a meter, and how to stop it happening again.

0 dBFSthe hard ceiling
−1 dBTPsafe export peak
3ways to detect it
−14 LUFSloud, not clipped

By Thomas, founder of CutScore · Updated June 2026

AUDIO CHECK · interview_take2.mp4
Two people leaning over a lit audio mixing board, watching the level meters during a recording, the exact moment where a signal can be pushed too hot and start to clip.
CRAFT SCORE
FIXES ADVISED
peaks hitting the ceiling
Clipping on loud words · peak +0.4 dBTP00:42
Loudness on target · −14 LUFS
Crackle audible · re-record this line00:42
The 30-second answer You know your audio is clipping when three things line up. First, the waveform has flat tops where loud peaks should be rounded, because the signal hit the ceiling and got chopped off. Second, you can hear a crackle, fuzz or harsh buzz on the loudest words, most obvious on headphones. Third, a meter shows the level touching or passing 0 dBFS, often lighting up red. Any one of those is a warning; all three together is clipping for sure. The lasting fix is to record with headroom and keep your final true peak at or below −1 dBTP. If checking by hand sounds tedious, that is the exact job CutScore does in one pass.
WHAT CLIPPING ACTUALLY IS

Picture pouring water into a glass that is already full. Past the rim, the extra water does not stack higher, it just spills and makes a mess. Digital audio works the same way. Every recording has a hard ceiling, called 0 dBFS, and your sound wave can ride right up to it but not above. When a loud moment tries to go higher, the system has nowhere to put it. So it slices the top of the wave off flat. That flat top is clipping, and it sounds like grit.

The frustrating part is that clipping usually happens on the moments you care about most. A laugh, a punchy word, a sudden loud action. Quiet speech behaves itself and then one excited sentence smashes into the ceiling and turns to fuzz. I have shipped this exact mistake more than once: a perfectly calm intro, then I get animated thirty seconds in, and that is the bit that crackles.

Here is the thing worth getting straight early: clipping is not the same as loud. Loud is a feeling. Clipping is a failure. You can have a video that is plenty loud, sitting near −14 LUFS, with not a single clipped peak. The targets are short and boring, and they are entirely under your control. Let us go find the clipping.

THE THREE TESTS

Three ways to know if your audio is clipping.

Eyes, ears and a meter. Each one catches clipping on its own, but when all three agree you can stop second-guessing yourself and fix it.

TestWhat clipping looks likeHow sure it makes you
Look · the waveformflat topsLoud peaks chopped into flat plateaus instead of rounded curves. Very reliable.
Listen · headphonescrackle & fuzzA harsh buzz or grit on the loudest words. Your ears rarely miss real clipping.
Measure · the meter≥ 0 dBFS, redThe level touches or passes zero and the meter lights up red. Exact and honest.
True peak on exportover −1 dBTPA true-peak meter reads above −1, so the platform re-encode may push it to clip.
Clip indicator lightstays litMany recorders and mixers have a clip LED that latches on once it overloads.
The sneaky caseSometimes the waveform looks fine but you still hear grit. That is often true-peak clipping introduced after export, when the platform converts your file and a peak that sat right at 0 dBFS sneaks over. This is why the safe target is −1 dBTP, not 0.
SKIP THE MANUAL CHECK

Scrubbing the whole timeline for one clipped word gets old fast. CutScore scans the entire track, flags every clip with a timestamp, and tells you whether it is fixable or a re-record.

Join the waitlist
HOW TO RUN EACH TEST

Catch the clip by eye, ear and meter.

1. The eye test: read the waveform

Open your clip in any editor and zoom in on a loud part. A healthy wave looks like rolling hills, peaks that rise and fall with rounded tops. Clipped audio looks like someone took a ruler to it: the tops are sliced flat, sitting in a straight line right at the edge. Those flat plateaus are the deleted peaks. The wider the flat section, the worse the clip. This one trick catches most cases in seconds, and it works in any timeline, free or paid.

2. The ear test: listen on the worst gear you own

Put on headphones and play the loudest moments. Clipping has a signature sound: a crackle, a fizz, or a harsh buzzy edge that arrives only on the loud words and vanishes when things go quiet. If clean audio is a smooth voice, clipped audio is that voice with sandpaper on it. Laptop speakers hide this, which is exactly why I trust them least. Headphones or decent monitors will expose grit your built-in speakers politely covered up.

An audio engineer wearing headphones and watching the input meters on a recording interface, the level check that catches a signal going too hot before it clips.
Headphones and a meter catch clipping that laptop speakers hide. Photo: Los Muertos Crew / Pexels.

3. The meter test: watch for red at zero

Every editor has a level meter, usually a vertical bar that jumps with the sound. The scale tops out at 0 dBFS, and that zero is the ceiling. When a peak hits or passes it, the meter flashes red and often holds a little marker at the top so you can see you overloaded even after the moment has passed. A peak indicator that latches red is the clearest yes there is. For the final word, add a true peak meter and keep it at or below −1 dBTP, because that catches the clipping that only appears after the platform re-encodes your file.

Loud is allowed. Clipped is not.

It is easy to confuse the two, so let me draw the line clearly. Loudness is how strong the whole track feels, measured in LUFS, and the YouTube target sits around −14 LUFS. You can hit that target and still be perfectly clean. Clipping is a specific peak failure: the wave tried to go past 0 dBFS and got cut. So you can be loud without clipping, and you can also clip a quiet video if your input gain was set far too hot. Aim for loud and clean, not loud at any cost.

RATHER SEE IT THAN READ IT?

Here is a real CutScore coaching report for an everyday video: loudness, true peak and clipping all scored, with the exact timestamps where the audio goes wrong.

See a sample report
FOUND CLIPPING?

What to do once you have spotted it.

Whether you can save the take depends on how hard it clipped. Work through these three in order, from prevention to rescue.

1
BEST FIXAUDIO
Re-record the clipped line with headroom
Hard clipping deletes the original peaks, so no plugin can truly bring them back. If you can re-record, drop your input gain so the loudest moments land around −12 to −6 dBFS, then deliver the line the same way. The clip disappears because the peak now has room to breathe.
How Lower the mic input or gain knob, watch the meter, and keep your loudest words clear of zero by several decibels.
2
SALVAGEAUDIO
Run a declipper on light clipping
If re-recording is impossible and the clipping is mild, a declipper plugin can rebuild the chopped peaks well enough to pass. It will not be perfect, but it often turns "obviously broken" into "fine." Pair it with a gentle cut to the harsh high frequencies and most listeners will never notice.
How Apply a declip or de-distort tool, then a touch of de-essing or a high-shelf cut to tame the leftover fizz.
3
PREVENTEXPORT
Cap your true peak at −1 dBTP on export
Even clean audio can clip after upload, because platforms re-encode your file and the conversion nudges peaks upward. A true-peak limiter set to −1 dBTP gives you the one decibel of safety that absorbs that nudge. It is the cheapest insurance against clipping you did not cause yourself.
How Add a true-peak limiter to your master, ceiling −1 dBTP, and confirm the export reads clean.
THREE WAYS TO CHECK

By ear, by meter, or in one pass.

OPTION 01

By eye and ear

Free, and good for obvious clips. Zoom the waveform, look for flat tops, then listen to the loud parts on headphones. The catch is that subtle true-peak clipping after upload can slip past your ears, and you have to remember to do it on every single video you make.

OPTION 02

With a true-peak meter

Accurate and honest. A loudness meter plus a true-peak meter tells you exactly where you stand against −14 LUFS and −1 dBTP. The cost is time and know-how: you open the tools, read them right, and do it for every export. Great if you like meters. Plenty of people do not.

OPTION 03

With a coach in one pass

Hand the file (or a link) to CutScore. It scans the whole track for clipping and true-peak overshoots, scores loudness against the right target, and gives you timestamps and fixes. No meters to read. See a sample report.

How CutScore catches clipping for you CutScore is an AI video quality coach for pre-publish QC. It measures the audio deterministically with an EBU R128 loudness meter and a true-peak meter, so clipping and overshoots are caught by maths, not by mood. You get one score from 0 to 100, the timestamp of every clipped moment, and whether each one is a quick fix or a re-record, before anyone else hears it. It judges the craft of the video itself, not your tags or thumbnails, so it sits next to a growth tool rather than competing with one. More on the method and the standards.
QUESTIONS

Frequently asked.

Look at the waveform first. Clipped audio has flat tops where the loud peaks should be rounded, because the signal hit the ceiling and got chopped off. Then listen on headphones: clipping sounds like a crackle, fuzz or harsh buzz on the loudest words. On a meter, anything touching or going past 0 dBFS is clipping. If all three agree, it is clipping.
Sometimes, not always. Light clipping can be softened with a declipper plugin or by reducing the harsh frequencies, and the result is usually passable. Hard clipping that lasted a while is gone for good, because the original peaks were deleted, not turned down. The honest fix is to re-record that line at a lower input level.
Loud is fine. Clipping is loud past the point the system can represent, so the tops of the wave get cut flat and turn to distortion. You can have a loud, perfectly clean video that sits near −14 LUFS and never clips. Loudness is how strong it feels overall; clipping is a specific failure at the peaks.
Aim for true peak at or below −1 dBTP on your final export. While recording, keep your loudest moments around −12 to −6 dBFS so you leave headroom for surprises. That one decibel of safety below zero matters because platforms re-encode your file, and the conversion can push a peak that sat right at 0 dBFS over the edge into clipping.
EARLY ACCESS

Stop the crackle before you publish.

CutScore scans your whole track for clipping and true-peak overshoots, then tells you exactly what to fix and where. Join the waitlist for early access.

Join the waitlist