How to normalize audio volume
Add audio files and choose targets; measurement and normalization run automatically, and downloads remain explicit. Each report includes integrated loudness, true peak, loudness range (LRA), gain change, and a clipping risk estimate.
Select audio files and set targets
Select MP3, WAV, FLAC, AAC, OGG, M4A, and more from the drop zone. Drag and drop is supported. Clicking browse files opens a picker using the browser's audio/* filter.
- Choose Loudness standard: EBU R128 (-23 LUFS) - EU broadcast delivery, ATSC A/85 (-24 LKFS) - US broadcast delivery, Podcast delivery spec (-16 LUFS), or Custom target....
- Choose a True peak ceiling preset or Custom....
- If you select Custom target... or Custom..., enter a valid numeric value before measuring. Invalid values are reported inline before processing starts.
- Choose Output format: WAV (16-bit PCM, recommended), FLAC (lossless, smaller file), or MP3 320 kbps (lossy re-encode). The available FFmpeg settings are applied during export.
These loudness presets are named delivery specifications, not universal streaming targets. Playback platforms may apply their own normalization regardless of the file's measured loudness.
Measure loudness
File cards appear when you add files. Cards update automatically with a measurement report after files are added or targets change. The report displays integrated loudness in LUFS, plus true peak in dBTP, loudness range in LU, gain change in dB, and a Low, Medium, or High clipping-risk estimate.
Changing the loudness target or true-peak target automatically remeasures and re-exports the current files. A failed remeasurement can leave an earlier measurement in the entry's data. Changing Output format does not require remeasurement.
Normalize and download the files
- After measurement completes, normalization runs automatically for entries with a valid report.
- Wait while the measured entries are processed sequentially. Entries without a stored measurement are skipped. Files are normalized separately; they are not merged.
- Download each completed file from its card, or use Download all (.zip). Multiple completed files are packaged as a ZIP; one completed file downloads directly.
The current Output format is applied automatically whenever processing runs. Use the icon button titled Remove in a file card's header to remove an item from the batch.
Output formats and WAV behavior
Outputs are decoded and re-encoded by FFmpeg. WAV uses pcm_s16le, so it creates a new 16-bit PCM WAV file. It is not a byte-preserving copy and may not retain a higher source bit depth. FLAC uses the FLAC encoder, while MP3 uses libmp3lame at 320 kbps.
Processing scope and errors
Audio processing runs in the browser through an FFmpeg/WebAssembly engine. The tool loads an approximately 30 MB engine before processing. This describes the audio-processing step only; it is not a claim about every request made by the page.
The implementation does not define a per-file size limit, total batch-size limit, or maximum file count.
Unsupported or unreadable files can show an Error status, and an individual measurement or normalization can fail. If needed, use Remove, add the file again, and measure it again.
Frequently asked questions
How does automatic processing work?
The tool measures each usable file and then normalizes it automatically. Changing a loudness or true-peak target starts a fresh current run before export. Changing the Output format does not require remeasurement.
What does true peak protect against?
True peak accounts for reconstructed levels between stored samples. The tool uses the predicted peak headroom to label clipping risk as Low, Medium, or High against your selected ceiling.
Will this trim silence or cut my audio?
No. It normalizes loudness and exports a new file, but it does not remove, trim, cut, or reorder audio content.
Does normalizing multiple files merge them?
No. Entries with a stored measurement are normalized independently and stay separate. An entry that cannot be measured may show an Error status. If a remeasurement fails, an earlier measurement may remain available; remove and re-add that entry before normalizing if you want to discard it.
Which input files can I select?
The drop zone lists MP3, WAV, FLAC, AAC, OGG, M4A, and more. You can drag files onto the drop zone or use browse files, which opens the browser's audio/* filter.
What happens if I use a custom target?
Enter a valid numeric value after selecting Custom target... or Custom.... Invalid values are reported inline before processing starts.
How is WAV output encoded?
WAV output is encoded as 16-bit PCM with pcm_s16le. It is a new file, not a byte-preserving copy, and may not preserve a higher source bit depth.
Is there a file-size or file-count limit?
The implementation does not define a per-file size limit, total batch-size limit, or maximum file count. It does load an approximately 30 MB processing engine.
Why is there a ~30 MB download the first time I use this?
Measurement and normalization run through a WebAssembly build of FFmpeg in your browser. The tool loads an approximately 30 MB engine before processing; persistent caching and instant future loading are not guaranteed.