In this guide
Normalize sets a sample-peak target; Loudness Normalization adjusts toward a measured loudness target. Choose the effect for the delivery requirement, preserve stereo balance, and measure the finished export for both loudness and peaks.
To normalize audio in Audacity, select the intended audio and open Effect → Volume and Compression. Choose Normalize to set a peak level, or Loudness Normalization to adjust toward a perceived-loudness target. Export and check the finished file: reaching one target does not automatically guarantee the other.
Keep the original and finish major edits before setting a final level. Neither effect repairs clipping already recorded or removes background noise; raising a quiet file also raises its noise.
Normalize or Loudness Normalization?
| Need | Audacity effect | What it does not guarantee |
|---|---|---|
| Set the highest sample peak | Normalize adjusts gain toward a chosen peak in dB. | Different recordings with the same peak may still sound very different in loudness. |
| Target measured perceived loudness | Loudness Normalization can use a LUFS target. | It does not guarantee a safe true peak or even out every sentence. |
| Control large differences between words | Consider editing levels or carefully comparing compression before final normalization. | Normalization alone applies an overall gain change; it is not a compressor. |
Use a target supplied for your destination when one exists; a number chosen for a podcast is not automatically appropriate for broadcast or a game. The peak versus loudness guide explains why these measurements differ.
Set a peak level with Normalize
- Make a working copy. Select the full recording or the exact passage you intend to adjust. Check that there are no unintended tracks in the selection.
- Open Normalize. Choose Effect → Volume and Compression → Normalize. Enable peak normalization and enter the desired maximum peak; for a simple working example, −1 dB leaves some sample-peak headroom below 0 dB. Follow the actual delivery specification when it differs.
- Check stereo and DC options. Leaving “Normalize stereo channels independently” off applies the same gain to the two channels of a stereo track and preserves their relative balance. The separate DC-offset option corrects an offset; it is not a noise-reduction switch.
- Apply and listen. Check the loudest point and quiet passages. Reopen the exported file and measure it after any later processing or lossy encoding.
Audacity's Normalize manual documents these controls. It also notes that separately selected tracks can be normalized independently to the same peak, which can alter the intended balance of a multitrack mix. If the balance between tracks matters, prepare and check the complete mix instead of blindly normalizing each track.
Set a LUFS target with Loudness Normalization
- Select the intended complete program or a representative working mix, then choose Effect → Volume and Compression → Loudness Normalization.
- Select Perceived Loudness and enter the required LUFS value. Audacity also offers RMS; RMS and LUFS are different measurements, so select the one specified for your task.
- Review channel options. Leave independent stereo normalization off when the existing left/right balance is intentional. For mono, check “Treat mono as dual-mono” against the destination meter and delivery requirements; different mono conventions can produce different LUFS readings.
- Apply, export, then measure the finished audio and listen at a comfortable level. Inspect the loudest section and quiet speech as well as the integrated loudness.
Audacity's Loudness Normalization manual explains its LUFS, RMS, stereo and dual-mono options. An integrated target alone is not a peak ceiling: if the required gain would create peaks above the allowed range, revise the mix or use suitable peak control, then remeasure. The podcast loudness and true-peak guide walks through that distinction.
Common problems
- The result is still too quiet: a single transient may determine a peak-normalization setting. Inspect the recording and compare a suitable dynamics edit instead of assuming peak normalization will match perceived loudness.
- Clipping appears after making it louder: undo, reduce gain or revise the dynamics, then check the exported file. Lowering a previously clipped file does not reconstruct its missing peaks.
- The stereo image shifts: check whether the channels were processed independently and return to the original if the change was unintended.
- Another LUFS meter disagrees: verify that both measured the same full file, channel layout and mono convention. Check settings and meter limitations before adjusting again.
The Saturalabs loudness meter can inspect a supported mono or stereo MP3/WAV up to 50 MiB and five minutes locally in the browser. Check its stated measurement limitations and compare the actual downloaded file. For peak-only adjustment in the browser, the separate audio normalizer uses a sample-peak target rather than a LUFS target.
Should I normalize before or after compression?
Compression reacts to the level entering it, so changing gain before the compressor can change how much it works. Make and compare the dynamics edit first, then set the final level and verify the export. See Audacity compressor settings for a voice-focused starting workflow.