Why Is My Audio So Quiet? Decibels, Gain and Clipping Explained

You recorded the meeting, the lecture or the voice memo, and on playback you have to push your laptop to full volume just to make out the words. Then a notification arrives at normal level and nearly knocks you off your chair. Only your file is quiet.

Turn it up too far, though, and the voice starts to crackle. This article explains why your audio is so quiet, what turning it up does to the file, and how far you can push it before it breaks.

Why a recording ends up quiet

Digital audio has a hard ceiling. Every sample in the file is a number, and there is a largest number the format can hold. That ceiling is called 0 dBFS (decibels relative to full scale). Every level in a recording is measured as a distance below that ceiling, so a voice that peaks at −18 dBFS sits 18 decibels under the maximum.

A phone or laptop microphone doesn't know whether you're about to whisper or shout, so it leaves a lot of room at the top. Sitting far from the mic, a soft voice or a low input level push the voice further down. The result is technically perfect and practically inaudible on small speakers.

That gap between the loudest moment in the file and the 0 dBFS ceiling is called headroom. A quiet recording is, almost always, a recording with too much of it.

The teacher whose lecture students can't hear

A teacher records a 40-minute lecture on her laptop for students who missed class. She checks it that evening: she can hear herself, barely, with the laptop volume maxed. Next morning, three students report the recording is "broken": it plays silently on their phones. It isn't broken. Its loudest moment peaks around −15 dBFS, and her voice sits well below that.

What turning it up actually does

Raising the volume of a file is called applying gain, measured in decibels too. Gain multiplies every sample by the same amount, so quiet and loud parts move up together toward the ceiling. Decibels are logarithmic:

• +6 dB roughly doubles the size of the waveform.
• About +10 dB is what most listeners describe as "twice as loud."
• +3 dB is a small but noticeable lift.

The key point is that gain moves everything by the same amount. So the safe amount of gain for a given file is set by its single loudest moment, not by its average level. If the voice peaks at −15 dBFS, you can add up to 15 dB before the peak touches the ceiling.

Why it crackles when you push it too far

Go past that point and the samples that would land above 0 dBFS have nowhere to go. The file can't store a number bigger than its maximum, so those samples get cut off flat at the ceiling. That is clipping, and you hear it as crackle, buzz or a harsh fuzz on every loud syllable.

This is why one loud moment can ruin the job. A quiet lecture with one cough next to the mic might peak at −2 dBFS while the voice sits around −20. Plain gain can only add 2 dB before the cough clips.

The usual answer is a limiter. A limiter watches the signal and turns down only the moments that would cross the ceiling, leaving everything else alone. The voice gets its full boost, and the cough gets caught just under 0 dBFS instead of being sliced flat. Push a limiter very hard and the loud moments start to sound squashed, but for the occasional spike in a spoken recording it is the difference between usable and ruined.

Gain also can't tell the voice from the background: room hum and hiss come up by the same amount. A boost makes a quiet recording audible, not cleaner.

The podcaster whose episode turned to fuzz

A podcaster's guest sounded faint, so he dragged the volume in his editor all the way up and exported. Now every laugh breaks up into a crackly fuzz. The editor pushed the peaks past 0 dBFS and the file stored them flat. Turning that export back down just gives him a quieter crackle.

"Compression" means two different things in audio

Search for help with quiet audio and someone will tell you to "compress it." The word covers two unrelated ideas.

Dynamic range compression narrows the gap between the quiet and loud parts of a recording, which can make it sound louder overall. A limiter is an extreme, fast version of it.

File compression makes the file smaller, usually by lowering the bitrate of an MP3. It changes how much space the audio takes, not how loud it is.

If your file is quiet, you want gain, with a limiter catching the peaks. If your file is too big to email or upload, you want the other kind.

Making a quiet file louder, safely

The MyTools volume booster is built around the ideas above. When you add a file, it reads the peak level and shows it next to the file name, so you can see your headroom before you pick a number. You boost the audio volume with a gain slider, then press and hold Compare to hear the original against the boosted version. The limiter is on by default and holds the output just under 0 dBFS. Everything runs in your browser, so a private meeting or a lecture is never uploaded.

How the teacher gets her lecture back

The teacher drops the lecture into the booster and sees a peak of around −15 dB. That peak is the moment she knocked the desk; her voice sits lower. She sets the gain to +18 dB, leaves the limiter on, and listens to a minute with Compare. Her voice now plays at a normal level on a phone speaker, and the desk knock, which would have landed 3 dB over the ceiling, is held just under it instead of cracking. She shares the new file and the "broken recording" emails stop.

How the podcaster rescues the episode

The clipped export can't be repaired, because the tops of those waveforms are gone. So the podcaster goes back to the original recording of his guest and runs it through the booster with the limiter on, so the laughs are caught under 0 dBFS rather than sliced off. The guest comes up to match his own voice, without the fuzz.

And if the boosted episode is then too large for his host's upload limit, that's a job for the other kind of compression: he can compress the audio file to a lower bitrate without touching the level he just fixed.

The short version

Your audio is quiet because the recording left too much headroom under the 0 dBFS ceiling. The loudest peak decides how much gain you can add, and pushing past the ceiling causes clipping you can't undo. Boost from the original file and keep a limiter on to catch the peaks.

Ready to make that recording audible? Try Boost audio volume for free →

File too big after the boost? Compress audio → is right there too.