Volume Booster Online: Raise a Quiet File Without Hard Sample Clipping
A track with a peak at -12 dBFS has room for roughly 12 dB of linear gain before its loudest sample reaches the digital ceiling. That does not mean it needs 12 dB of gain. It may be quiet because it has wide dynamics, not because its entire waveform was saved at the wrong level. Load your recording above, start at a comfortable playback volume, and compare the original with a modest boost. The file stays on your device while you work.
Added gain: what the dB slider changes
The gain slider runs from 0 to +24 dB. Gain is a multiplication of sample amplitude. A +6 dB setting approximately doubles the signal amplitude before the compressor reacts; +12 dB is about four times the amplitude. This is not the same as doubling perceived loudness. Perception depends on frequency content, duration, and listening level. A quiet intro and a loud chorus will both rise, so the gap between them remains unless later dynamics processing changes it. If the loud part already sits close to full scale, adding gain can make the compressor work hard and may change the sound noticeably.
First try +3 to +6 dB on a recording that is only slightly quiet. Listen to the loudest part, not just the first few seconds. If drums lose impact, a backing track pumps, or the voice sounds squeezed, lower the gain. The browser's compressor has a -3 dBFS threshold, a 20:1 ratio, a 3 ms attack, and a 150 ms release. These values make it a safety stage for loud peaks, not a transparent true-peak limiter. A strong boost can still sound distorted if the input was already clipped or the compressor is forced to act throughout the recording.
When to use peak normalization
Peak normalization measures the largest sample in your decoded source. When enabled, it applies the gain needed to place that peak around 0.95 in linear amplitude, which is about -0.45 dBFS, before applying your added-gain setting. It is useful when a clean voice memo was recorded at a very low level. It is less useful when one isolated click is already loud; that click will determine the normalization amount even if the rest of the file is quiet. Normalization also raises any background hiss along with the music. It cannot recover lost detail from a noisy microphone or repair a damaged source.
If you enable Normalize and leave Added gain at +6 dB, both operations happen in sequence. That may drive the compressor much harder than normalization alone. Compare one change at a time. If your goal is a precise streaming loudness target, use the LUFS loudness meter after export to measure the result. This booster does not calculate integrated loudness and cannot promise a specific LUFS reading. The MDN GainNode reference explains the browser's gain stage, while the DynamicsCompressorNode reference explains the compressor that follows it.
Why the exported peak can be lower than expected
The browser renders the file offline at its decoded sample rate, which may differ from the source file's rate. After compression, the tool checks every output sample. If any sample exceeds 0.99, it scales all channels equally so the highest one is 0.99. That final safeguard prevents hard sample clipping in the 16-bit WAV encoder. It does not detect inter-sample peaks or guarantee how a downstream MP3 encoder will behave. If the file sounded loud during preview but exports with little improvement, try a smaller added-gain value. Too much gain can cause the compressor and peak safeguard to cancel out the apparent increase.
Take a two-minute phone recording that peaks at -18 dBFS. Start with Normalize on and Added gain at 0 dB. Listen for background noise and for changes in the loudest moments. If it still needs more body, raise Added gain gradually instead of jumping to +24 dB. For a track with big differences between quiet and loud sections, the dynamics compressor gives you direct control over threshold, attack, release, and makeup gain. A simple volume increase cannot selectively raise quiet words without affecting loud ones.
Files, formats, and the next step
The tool accepts mono or stereo browser-decodable audio up to 100 MB, with a 128 MB decoded-memory guard. The result is always uncompressed 16-bit PCM WAV, not an MP3, even when you start with an MP3. To estimate that output size, use the sample rate and bit depth calculator. If you need a smaller file, run the downloaded WAV through the audio converter afterward. Changing the volume of an MP3 requires decoding it first; re-encoding to MP3 later can add generation loss. Save the source so you can make another pass if necessary.
Before you share the result, listen from start to finish at a safe level. A single loud transient can hide problems that do not appear in a short preview. If you hear rough edges, compare against the original and reduce the boost instead of exporting another louder copy.
