How to use Audio Volume Adjuster
- 1Drop in an audio file — its peak level is measured as it loads
- 2Set a gain in decibels, or switch on normalise
- 3Watch for the clipping warning, which appears before you export
- 4Add a fade in or out if the level change makes the ends abrupt
- 5Export and download
What Audio Volume Adjuster does
Volume is measured in decibels, and decibels are logarithmic rather than linear. That is the single thing worth understanding before changing one: +6 dB is roughly twice the amplitude, +20 dB is ten times it, and −6 dB is half. A change of 1 dB is around the smallest most people notice at all.
The ceiling matters as much as the change. Digital audio has a hard maximum, and any sample pushed past it is not made louder — it is flattened, which is heard as distortion. Because the file's true peak is measured when it loads, this tool can tell you that a given gain will clip before you spend time exporting it.
When to use it
- A recording that is too quiet
- A voice memo made too far from the microphone can usually be lifted substantially. Normalise finds the maximum safe increase automatically.
- Matching tracks recorded separately
- Interviews recorded across several sessions vary in level. Normalising each to the same target makes them consistent before joining them.
- Taming something too loud
- Audio that distorts on playback is often simply too hot for the device. A few decibels down usually fixes it.
- Preparing a background bed
- Music under narration typically wants to sit 15 to 20 dB below the voice, which is a specific number rather than a guess once you can type it.
Audio Volume Adjuster settings explained
- Gain (dB)
- A direct adjustment applied to the whole file. Positive raises, negative lowers. Because decibels are logarithmic, +6 is about double the amplitude and −6 about half; 1 dB is near the threshold of noticeability.
- Normalise
- Instead of choosing a gain, you choose a target peak and the gain is calculated from the file's measured maximum. This finds the largest increase that does not clip, which is usually what people actually want.
- Target peak
- Where the loudest moment should land, defaulting to −1 dB. Leaving a decibel of headroom rather than going to 0 avoids the rounding errors that can cause clicks when a file is later re-encoded.
- Clipping warning
- Shown before export, predicted from the peak measured when the file loaded. It tells you the chosen gain will push samples past the maximum, so you can adjust rather than discovering the distortion afterwards.
- Fade in and out
- Ramps the level at the start and end, in seconds. Useful when a raised level makes the beginning of a recording start abruptly.
What decibel changes actually sound like
| Change | Amplitude | Perceived as |
|---|---|---|
| +20 dB | 10× larger | Dramatically louder |
| +6 dB | 2× larger | Clearly louder |
| +3 dB | 1.4× larger | Noticeably louder |
| +1 dB | 1.1× larger | Barely perceptible |
| −6 dB | Half | Clearly quieter |
Limits and what to expect
- Raising the level raises the noise floor with it — hiss and hum get louder too.
- Clipping already present in the source cannot be repaired; the flattened peaks are gone.
- This changes level, not dynamics. Quiet passages stay proportionally quiet — that would need a dynamic-range compressor.
- Normalising works from a single peak, so one brief spike limits how much the rest can be raised.
- 150 MB per file, and long recordings decode slowly.
Troubleshooting
- Normalise barely raised anything
- One momentary spike is setting the ceiling — a cough, a bump on the microphone, a door. Trim that moment out first, then normalise, and the rest of the recording can come up much further.
- It got louder but also hissier
- Gain is applied to everything, including the noise floor. If the recording has audible hiss, raising the voice raises the hiss by exactly the same amount. Noise reduction is a separate operation this tool does not perform.
- The clipping warning appeared
- Your gain would push the loudest samples past the digital maximum, where they are flattened rather than amplified. Reduce the gain, or use normalise, which calculates the largest safe increase for you.
- It is still too quiet even at the maximum
- The peak is already near the ceiling but the average is low, which means the recording is very dynamic. Raising the average without touching the peak needs a dynamic-range compressor, which changes loud and quiet parts differently.
Your files and your privacy
The peak measurement, the gain and the export all run inside your browser using its own audio engine. Nothing is uploaded and nothing is stored after the tab closes. The recording never leaves your device at any point, including while it is being played back for you to check.
What to use next
FAQ
How many decibels should I add?
Use normalise rather than guessing — it measures your file's peak and calculates the largest increase that will not clip. If you are setting it manually, remember the scale is logarithmic: +6 dB is roughly double the amplitude and +1 dB is close to imperceptible.
What does clipping mean?
Digital audio has a hard maximum. Samples pushed beyond it are not made louder, they are flattened to the ceiling, and that flattening is heard as distortion. It cannot be undone afterwards, which is why the warning appears before you export rather than after.
Why normalise to −1 dB instead of 0?
Leaving a decibel of headroom protects against the small rounding errors introduced when a file is later re-encoded to MP3 or AAC. A file peaking at exactly 0 dB can end up fractionally over after that conversion, producing clicks.
Why did normalising barely change anything?
Because it works from the single loudest sample. One brief spike — a cough, a knock on the microphone — sets the ceiling for the entire file. Trim the spike out and normalise again, and the rest can rise much further.
Can this make quiet parts louder without changing loud parts?
No. This applies one gain to the whole file, preserving the relationship between loud and quiet. Changing that relationship requires a dynamic-range compressor, which is a different kind of tool.
Is my audio uploaded?
No. Measuring, adjusting and exporting all happen inside your browser on your own device.