Audio · 05
Volume booster
Raise a quiet voice note to a sensible peak, or add a fixed boost. Peaks that would pass full scale are clipped so the file stays valid.
How to use it
Peak normalize finds the loudest sample and scales the file so that sample lands near −1 dB. That is the right move for a voice memo that was recorded too low and has no huge spikes.
+6 dB and +12 dB multiply every sample by a fixed gain. If the take is already loud, those peaks flatten. You will hear that as a harder edge, not as more detail.
The download is 16-bit WAV at the decoded sample rate. This is not a loudness-unit normalizer and it does not match two files to each other. Run it once, then stop.
Tips
- Normalize a quiet interview before you run the noise remover. RNNoise behaves better when the voice is not a whisper in the noise floor.
- Do not stack +12 dB twice. Clip once, then leave it.
- If the meter looks fine but the voice still feels quiet next to music, the issue is loudness, not peak. Mix that in your editor.
FAQ
Will this fix distortion that is already in the file?
No. Clipping that was recorded is still clipping. Gain only changes the level of what is there.
Why WAV and not MP3?
A level change should not add another lossy encode. Convert afterwards if you need a smaller file.
Is the boost uploaded for processing?
No. Web Audio decodes the file in the tab and writes the WAV there.
License
This page uses the browser Web Audio API. No model is downloaded, and the audio is not sent anywhere.