You can increase the volume of an audio file online by loading it into a browser tab, choosing a gain in decibels or a loudness target in LUFS, and exporting a louder copy. The processing runs locally with FFmpeg compiled to WebAssembly, so nothing is uploaded and no account is needed.
Two facts decide whether the result sounds better or worse. Gain multiplies the waveform that is already there, so it cannot bring back detail that was never recorded; it raises the noise floor along with the voice. Every +6 dB roughly doubles amplitude, and anything past about +6 dB can push peaks beyond 0 dB, where samples are clipped into a harsh crackle. That is why the tool ships an optional limiter, which holds peaks below full scale at the cost of slightly tighter dynamics, and why loudness normalization exists as a separate mode that corrects the whole track to a target level instead of applying one flat gain.
Quiet recordings are the most common audio complaint that has nothing to do with quality. A phone memo captured too far from the speaker, a lecture recorded on a laptop microphone, a download whose dialogue sits far below the music: the file itself is fine, it is simply too low to hear comfortably on a phone speaker. This guide covers what raising the level actually changes, when a flat gain is the right move and when loudness normalization is, and where the approach stops working.
Volume Booster
Raise or lower a track by up to 20 dB, or normalize it to a streaming, podcast or broadcast target. The FFmpeg filter chain runs in your browser.
What raising the volume actually changes
Volume is amplitude, and amplitude is not the same thing as information. When you raise the level, the tool multiplies every sample in the file by the same factor, so the quiet parts and the loud parts move up together. Nothing new appears: the recording still contains exactly what the microphone captured, only written back at a higher number on disk.
Louder is not clearer
Because the multiplication is uniform, the background hiss, the room rumble and the microphone self noise come up by the same amount as the voice. On a very quiet recording the noise is often what you notice first after boosting, which is why the useful range is narrower than the slider suggests. If a take sounds muffled or noisy, turning it up does not fix that; it only makes the problem easier to hear.
Why +6 dB is roughly the practical limit
Every +6 dB doubles the amplitude of the signal. A track that already peaks near full scale has no headroom left, so pushing it louder sends the tallest samples past the maximum the format can store, and those samples are flattened into a hard crackle rather than made louder. A source that peaks around -18 dB has room to move; one that peaks at -1 dB does not. The tool flags this as you push the slider past +6 dB and offers an anti-clipping switch, but neither option can invent level that the file does not contain.
How to increase the volume of an audio file
The whole job is three steps, and it runs in the browser tab with no install and no account. The tool opens MP3, WAV, M4A, AAC, OGG, FLAC and Opus files, one at a time, and reads them from your disk into page memory.
- Add your audio. Drop a file onto the upload box or click to browse. The file stays on your device, the panel lists its size and duration, and a file that is not audio is refused with a short note instead of being half-processed.
- Choose a mode and set the level. Stay on Boost by dB and drag the slider anywhere from -20 dB to +20 dB, or switch to Normalize loudness and pick a target. Turn on anti-clipping when the boost is large.
- Adjust volume and download. Click Adjust volume, listen to the preview in the result card, then press Download. The result card shows the source size next to the output size so you can see what the export produced.
The first click does download the FFmpeg transcoding engine, about 30 MB, which is cached in the browser afterwards. That download is code, not your audio, and after the first run the tool keeps working offline.
Boost by dB or normalize loudness
These two modes both make a track louder, but they answer different questions. One is a fixed multiplier that you set; the other measures the track first and then corrects it to a target level. Confusing them is the main reason people boost a file twice and end up with a distorted result.
| Mode | What it does | Level it aims for | Best for |
|---|---|---|---|
| Boost by dB | Multiplies the whole file by one fixed gain, from -20 dB to +20 dB, default +6 dB, with an optional limiter | Whatever the source becomes; the relative levels never change | A single track you simply need louder or quieter |
| Normalize loudness | Measures the track first, then corrects it to a target using EBU R128, holding true peak at -1.5 dBTP and loudness range at 11 LU | -14, -16 or -23 LUFS | Several clips that must sit at the same level |
Use a flat gain when you control the whole track
If the recording is one piece and the only problem is that it is too quiet, a fixed gain is the honest answer. You hear exactly what you set: +6 dB in the preview is +6 dB in the download, and the internal balance of the take is untouched. It is also the faster mode, since nothing has to be analysed before the export starts.
Use loudness normalization for episodes and uploads
Loudness normalization follows the EBU R128 and ITU-R BS.1770 standard that streaming and podcast platforms measure against. The tool offers Streaming at -14 LUFS for Spotify and YouTube, Podcast at -16 LUFS for spoken word, and Broadcast at -23 LUFS for TV delivery, and it holds the true peak at -1.5 dBTP and the loudness range at 11 LU. Reach for it when several recordings have to sit at the same level, because it corrects each one to the same target instead of preserving their mismatch.
Anti-clipping, and what the limiter costs you
Anti-clipping adds a limiter stage after the gain that holds the peaks at 0.97 of full scale, just below the maximum a sample can store. Without it, any sample that would run past zero is cut off flat, and the ear hears that as a crackle on the loudest moments. Turning it on is the difference between a loud file and a loud file that still sounds clean.
There is a trade-off. A limiter reduces the difference between the loud and quiet parts, so a heavily limited track sounds a little flatter and loses some of its punch. For a gentle boost on a recording with plenty of headroom, it is usually fine to leave the switch off. For anything past about +6 dB, or a source that already sounds hot, turn it on.
Output formats and file names
Every volume change is written into a new file, so the result is re-encoded in the format you pick rather than copied. The tool exports MP3 at 192 kbps for sharing, uncompressed WAV when the file goes back into an editor, and M4A at 192 kbps for a smaller file at similar quality. The quality therefore depends on that choice rather than on the volume step itself, and your original file is never modified.
The download is named after your source with a suffix that records which mode ran. A decibel gain produces a name ending in -volume, and a normalization pass produces one ending in -loud, so a boost and a normalize of the same track never overwrite each other. A practical habit is to keep the working copy in WAV while you are still editing, then export MP3 or M4A once, at the end, for whatever leaves your machine.
What this tool cannot do
The limits are worth reading before you commit, because in these cases a different tool is genuinely the right answer.
- Boosting cannot restore lost detail. Gain multiplies what is already in the file, so muffled speech, clipped peaks and missing high frequencies stay exactly as they were. It also raises the noise floor with the voice, so a very quiet recording can end up sounding hissier after a large boost.
- No noise reduction, no equalizer, no channel or vocal separation. The filter chain is gain, an optional limiter and an optional loudness correction, and nothing else. Cleaning up a noisy or muddy recording needs a tool built for that job.
- The limiter compresses dynamics. It removes clipping by holding peaks down, which also reduces the contrast between the loud and quiet sections of the track.
- Everything is re-encoded. A volume change cannot be written losslessly into the same container, so the output is encoded again, and the bitrate and quality depend on the format you choose.
- The first run downloads about 30 MB. That is the FFmpeg engine, cached afterwards, and on a slow connection the first click sits at the loading stage for a while.
- One file at a time. There is no batch queue, so a set of clips is processed one after another rather than in a single pass.
- Long recordings depend on device memory. No duration limit is written into the tool, but the engine runs in browser memory, so a very long file can fail on a device with little free memory.