How many decibels should I add?

Start with 6, which sounds like roughly double. If you cannot judge it, use the target-level mode instead — it measures your file and works out the number, which is more reliable than guessing twice.

Why did adding more gain stop making it louder?

Because the loudest moments have reached the maximum the format can hold. Beyond that the peaks are held flat rather than made bigger, which changes the shape of the sound without raising it. The result panel warns you when this is happening.

Is this the same as compression or loudness normalisation?

No. This scales everything by one factor, or scales it so the peak lands on a target. Broadcast loudness normalisation measures perceived loudness over time and squeezes the dynamic range, which is a different and much more invasive operation.

Can I make a recording quieter?

Yes — use a negative amount. Reducing gain is entirely safe: there is no floor to hit the way there is a ceiling above.

Change Volume

Make a recording louder or quieter, or level it to a target.

Everything runs inside your browser. Your files never leave your device.

Drop your files here

.mp3, .wav, .m4a, .aac, .ogg, .oga, .opus, .flac, .webm, .weba · they stay on this device

Ask about Change Volume

Questions about what this tool does, which option to pick, or what it can and cannot handle.

This is the one part of the site that uses a server. The question you type here is sent to an AI provider to be answered — your files and whatever you put in the tool above are not, and the assistant cannot see them. Answers are generated and can be wrong; the tool itself is not guessing.

About the Change Volume

There are two entirely different problems that both get described as "the volume is wrong", and they need different answers.

The first is that a recording is simply too quiet or too loud, and you know roughly by how much. That wants a straight change in gain: every sample multiplied by the same number. Six decibels up is roughly twice as loud, six down roughly half, and the relationship between the loud and quiet parts is untouched.

The second is that you have no idea by how much — you just want it to sound like everything else. That wants levelling to a target: the tool finds the loudest moment in the file, works out what would put that moment just below the maximum, and applies that. A voice note recorded at arm's length and a podcast recorded into a proper microphone end up at comparable levels without you guessing.

The ceiling matters, and this is where cheap tools do damage. Digital audio has a hard maximum. Push a sample past it and, written as an integer, it wraps around to a large negative number — a peak becomes a burst of noise, and a file made "louder" comes back ruined. Samples are clamped at the maximum here instead, so an over-enthusiastic setting flattens the loudest peaks rather than shattering them. When that starts happening the result panel says so, and points at the levelling mode, which cannot overshoot by construction.

The default target is one decibel below maximum. That is the convention: as loud as the format allows with a sliver of room left, so that a player's own processing has somewhere to go.

None of this is destructive to your original. The file on your disk is untouched; a new one is written. And nothing is uploaded — the arithmetic happens in this tab, on samples the browser decoded there.

How to use it

1Drop in one recording or a batch of them.
2Choose whether to change by a set amount or to match a target level.
3Set the amount, or the level the loudest point should reach.
4Run it, then listen to the result before you rely on it.

Questions

How many decibels should I add?: Start with 6, which sounds like roughly double. If you cannot judge it, use the target-level mode instead — it measures your file and works out the number, which is more reliable than guessing twice.
Why did adding more gain stop making it louder?: Because the loudest moments have reached the maximum the format can hold. Beyond that the peaks are held flat rather than made bigger, which changes the shape of the sound without raising it. The result panel warns you when this is happening.
Is this the same as compression or loudness normalisation?: No. This scales everything by one factor, or scales it so the peak lands on a target. Broadcast loudness normalisation measures perceived loudness over time and squeezes the dynamic range, which is a different and much more invasive operation.
Can I make a recording quieter?: Yes — use a negative amount. Reducing gain is entirely safe: there is no floor to hit the way there is a ceiling above.