A quiet recording can be inconvenient to hear, but raising every sample also raises background noise. Audio Volume Booster applies a chosen gain multiplier and peak limiting, then exports an MP3. It is a practical level adjustment rather than a restoration or noise-removal process.
Begin with a modest multiplier
Choose one supported audio file no larger than 50 MB. Enter Volume multiplier between 0 and 4. A value of 1 requests unchanged gain, values above 1 increase it, and values below 1 reduce it. The default 1.5 is a moderate first comparison.
The processing chain applies gain and then a peak limiter set to 0.95. This helps control peaks after amplification, but it does not establish a broadcast loudness target or make every sentence equally loud. It also cannot restore detail already lost to clipping in the source.
Select Process media and keep the browser tab open. The export is a 192 kbps MP3. First use downloads a roughly 32 MB single-thread WebAssembly engine, and the work runs locally in your browser rather than uploading the recording to ERITHRA.
Compare gain on a simple example
Suppose a quiet passage contains samples with an amplitude of 0.2. A multiplier of 2 requests 0.4 before the limiter. A sample at 0.6 would request 1.2, so the limiter becomes relevant for that louder peak. These sample calculations describe the gain step, not a guarantee of perceived loudness.
Export at 1.5 first and listen at a consistent device playback volume. Compare speech clarity, background hiss, and abrupt loud sounds. If the source already has harsh peaks, a larger multiplier may make the result less pleasant even though it sounds louder.
Use the original source for each trial rather than repeatedly boosting a previously exported MP3. This keeps the comparison clearer and avoids unnecessary generations of lossy re-encoding.
Know when to use another workflow
A recording with wildly different speaker levels may need detailed editing or loudness analysis. This tool has no per-speaker adjustment, denoising control, or automated silence editing. Multiplier 0 intentionally silences the audio; it is not a restore-default setting.
Processing can use considerable memory and has a two-minute execution limit. Unsupported codecs or damaged sources may fail; Cancel processing can stop a job. Review metadata and playback before sharing, and keep the original. Use Audio Trimmer to isolate the relevant passage first, or Audio Format Converter when the next application needs another supported audio format.