Volume Booster
Drop in a file, choose how many decibels louder or quieter it should be, and export a WAV. It shows the peak and average level before and after, handles files up to 150 MB and 15 minutes, and nothing is sent anywhere.
Use audio you made or have permission to use. Nothing is sent anywhere, so we never hear or keep it.
Level set. Melodious can turn the finished track into a music video — See the music video maker
Six worked examples
Each case is the tool's own arithmetic run on a generated test tone, half a second long at 44,100 Hz. Peak is the loudest single sample; RMS is the average level. Both are in decibels relative to full scale, where 0 dBFS is the loudest a file can be.
A quiet recording, +6 dB, never clip
- Before: peak -20 dBFS, RMS -23 dBFS
- Gain applied: +6.00 dB (asked for +6 dB)
- After: peak -14 dBFS, RMS -17 dBFS
- Samples bent or cut flat: 0
A louder one, +12 dB, never clip holds the gain back
- Before: peak -6 dBFS, RMS -9 dBFS
- Gain applied: +5.92 dB (asked for +12 dB)
- After: peak -0.1 dBFS, RMS -3.1 dBFS
- Samples bent or cut flat: 0
The same +12 dB with the soft limit
- Before: peak -6 dBFS, RMS -9 dBFS
- Gain applied: +12.00 dB (asked for +12 dB)
- After: peak -0.1 dBFS, RMS -1.2 dBFS
- Samples bent or cut flat: 15,540
The same +12 dB with a hard clip
- Before: peak -6 dBFS, RMS -9 dBFS
- Gain applied: +12.00 dB (asked for +12 dB)
- After: peak 0 dBFS, RMS -1.1 dBFS
- Samples bent or cut flat: 14,660
Turning it down, -6 dB
- Before: peak -6 dBFS, RMS -9 dBFS
- Gain applied: -6.00 dB (asked for -6 dB)
- After: peak -12 dBFS, RMS -15 dBFS
- Samples bent or cut flat: 0
Asking for more than the tool allows is refused
- Refused: Choose a change between -40 and +40 dB.
The widget reads the same two numbers back from the WAV it writes, so you can compare the before and after of your own file the same way.
How to make audio louder without ruining it
Drop an audio file onto the box. The page decodes it and reports two numbers: the loudest single sample, called the peak, and the average level over the whole file, called RMS. Both are in decibels relative to full scale, so 0 dBFS is the loudest a file can be and every ordinary level is a negative number.
The gap between your peak and 0 dBFS is your headroom. A recording with a peak at -12 dBFS has about 12 dB of room to grow before anything is lost. Enter a gain up to that figure and the file gets louder with the shape of every wave preserved. Beyond it, something has to give, and the three modes below are the three ways of deciding what.
Never clip, soft limit or hard clip
Never clip is the default. It reduces the gain you asked for if the loudest sample would pass -0.1 dBFS, applies what fits and tells you the number it used. Nothing is bent, nothing is cut, and the page reads back from the written file that no sample sits at full scale.
Soft limit leaves everything below -1 dBFS untouched and bends only the peaks above it toward -0.1 dBFS, rounding them instead of slicing them. Hard clip applies the gain and flattens whatever passes the top. Both can make a file louder than Never clip allows, and both change the loudest parts for good, so the page counts the samples they altered and shows it next to the result. If that count is large, listen for harshness before you use the file.
Decibels, amplitude and what you actually hear
A decibel is a ratio, not a unit of loudness. Adding 6 dB multiplies every sample by about 2, adding 20 dB multiplies it by 10, and taking away 6 dB halves it. The page shows the percentage next to the box so the number means something. The ear is less direct: most listeners judge a sound as roughly twice as loud after an increase of about 10 dB.
This is also why a small boost often disappoints. Three decibels is about 1.4 times the amplitude, and it is hard to notice outside a side-by-side comparison. If you cannot hear the difference, the Play the result button and the before and after readout are the way to check you are not fooling yourself.
What this does not do
It applies one gain to the whole file, so it cannot make the quiet parts louder than the loud parts the way a compressor can. It does not measure loudness in LUFS or match a streaming target, and it does not remove noise. It handles mono and stereo, keeps the first two channels of anything larger, and writes a 16-bit WAV at the rate your browser decoded the file at. The limits are 150 MB and 15 minutes in, and 100 MB out.
For a recording that is simply too quiet, a gain is the right fix. For a mix that is quiet next to other records, the fix is in the mix and the master, and a volume booster would only trade away headroom you need.
Other things to do to the same file
If the file has dead air at the start or a cough in the middle, the audio cutter takes it out and can fade the ends. To put several recordings in one file at matching levels, boost each one here and then use the audio joiner. The reverse audio page plays a file backwards.
To see what a track looks like at the new level, the audio visualizer turns it into a moving picture, and the BPM finder reads its tempo.
From a track at the right level to a finished video
A video is judged by its sound as much as its picture, and a quiet track makes both feel weaker. For the wider route, the guide to making a music video for your song covers it end to end.
When the track should have scenes, a character and a story, the AI music video generator storyboards it and renders it to your audio. It is paid, from $19/month, and you see the cost before anything renders. The WAV you make here stays on your device unless you upload it yourself.
Frequently asked questions
How do I make an audio file louder?
Drop the file onto the box, enter how many decibels to add (3 to 6 is a clear step up), and leave Never clip selected. The page shows the current peak and average level, tells you the gain it will really apply, and lets you listen with Play the result. Then export the WAV and compare the levels it reads back from the file.
How many decibels is twice as loud?
Two different things get called loudness. Adding 6 dB doubles the amplitude of the wave, which is what the number on this page changes. The ear judges a sound twice as loud after roughly 10 dB. So +6 dB is clearly louder, +10 dB sounds about double, and anything near +20 dB is a different file, usually with problems.
Why did the tool apply less gain than I asked for?
In Never clip mode, the gain is capped so the loudest sample ends at -0.1 dBFS, just under the top of the scale. A file whose loudest peak is already at -3 dBFS has only 2.9 dB of room, so a request for +6 dB becomes +2.9 dB and the page says so. Choose the soft limit or the hard clip if you want the full amount and accept the cost.
What is clipping, and why does it sound bad?
A WAV can only store samples up to a fixed ceiling. If a boosted wave would go past it, the top is cut off flat, and a flattened peak is a harsh, buzzy distortion that cannot be undone. The hard clip option does exactly that and reports how many samples it flattened, so you know how much damage was done.
What does the soft limit do?
Samples below -1 dBFS are left exactly as they are. Above that, the curve bends gradually toward a ceiling of -0.1 dBFS, so a loud peak is rounded instead of sliced. The result is louder than Never clip allows and gentler than a hard clip, but it still changes the shape of the loudest parts. The page reports how many samples it bent.
Is this the same as normalizing or mastering?
No. This applies one fixed gain to the whole file. It does not measure loudness in LUFS, aim for a streaming target, compress the dynamics or equalize anything. If a track is quiet because it was recorded low, a gain is the right tool. If it is quiet next to commercial releases, that is a mastering job this page does not do.
Can I use it to make a quiet voice recording easier to hear?
Yes, for a recording that is quiet but clean. A gain raises everything equally, including background hiss and room noise, so a recording with a lot of noise gets a louder hiss. A good test is to listen to the pauses between words after boosting: if the noise is bothering you there, it will bother your listeners at the new volume.
What range of change does it allow?
From -40 dB to +40 dB in half-decibel steps on the slider, and any number between those on the box. It opens files up to 150 MB and 15 minutes and writes a WAV up to 100 MB, about 9.9 minutes of stereo at 44,100 Hz. Past a limit it refuses and says what would fit.
Does the booster upload or store my audio?
No. The file is read in your browser, multiplied by the gain in your browser and written back to your device. The page makes no request with your audio, and the code that applies the gain contains no network call. Closing the tab clears it. The levels you see are measured on your own device.
Is it fine to boost a track I did not record?
Only with permission. Use audio you made or have permission to use. Raising the level of a commercial recording and posting it can still infringe copyright, and the tool cannot tell whose audio a file holds, so the check is yours.
A track at the right level is ready to be seen
Attach it and describe the look. An AI director storyboards the song scene by scene and renders it to your audio. Plans from $19/month, and you see the cost before anything renders.
Start your videoMore tools: Audio cutter · Audio joiner · Reverse audio · Pitch changer · Music speed changer · Audio loop maker · BPM finder · Audio visualizer