Audio Normalizer

This implementation measures RMS energy and sample peak. It does not calculate integrated LUFS, loudness range or inter-sample true peaks. The term normalization describes its gain adjustment, not a certification for a broadcaster or streaming platform. Use the destination’s required meter when those standards matter.

A high peak may prevent the recording from being amplified enough to reach the average target without clipping. The tool reduces the requested gain to respect a -1 dBFS sample-peak ceiling. The displayed actual RMS explains the result. Compression or limiting would be a different process.

No. A file containing only zero-valued samples has no recorded information to amplify. The processor rejects that case rather than inventing sound. If you expected speech in a silent file, inspect the recording and export settings in the application that created it.

No. One gain is applied to the whole signal, including noise and existing distortion. A quiet noisy file becomes a louder noisy file unless peak protection restricts the gain. Repair or re-record a damaged source when possible.

No. The measurements consider the recording’s channels together and apply one gain to both. This preserves the relative stereo balance. It does not correct a recording where one microphone channel was much quieter than the other.

The operation changes amplitude, not playback speed or the number of samples. The file keeps its decoded duration and sample rate, then exports as WAV. Listen to the beginning and end as well as the level before replacing a source in another project.

RMS describes average signal energy, while LUFS measurement uses a different loudness model and measurement rules. This tool adjusts toward an RMS target with peak protection. It should not be described as a LUFS meter, true-peak limiter or certified broadcast mastering service. If a podcast platform, station or distributor specifies LUFS, use a suitable loudness meter and verify the final file against that requirement in your delivery workflow. An RMS value cannot simply be relabeled as LUFS, even if two settings happen to look numerically similar.

A shared target can be a useful starting point for similar material, but it is not a guarantee that different recordings will sound equally loud. Speech, sustained music and short impacts have different energy patterns. Compare the actual sequence and use your destination’s requirements where applicable. The normalizer applies one overall gain rather than riding every word or compressing individual peaks. If one sentence remains much quieter than another inside the same recording, you may need a more detailed editing or dynamics workflow instead of repeatedly normalizing the entire file.

Repeating the operation does not recover lost detail or make a recording progressively cleaner. Once a file has been adjusted, another pass simply evaluates that result and applies another gain change within the available limits. Work from an original copy, choose a sensible target and compare before and after. Avoid exporting a long chain of revisions when you can return to the original recording instead. If the result still has an obvious problem, identify whether it is a level mismatch, distortion, background noise or uneven speech; only the first of those may be solved by a simple overall gain.

Download the processed WAV and keep the original recording until you have checked the result in context. Note the target you chose and the displayed before-and-after measurements if you need to explain the change to a collaborator. The browser edit is temporary and is not automatically added to My creations. A file must fit the input limits of 25 MB and 120 seconds and be decodable by your browser. When you pass the result on, describe it accurately as an RMS-adjusted file rather than claiming that it has passed a broadcast loudness standard.