In this guide
Select the audio clips on the timeline, right-click and choose Normalize Audio Levels. Choose a normalization mode and target, decide whether clips should be adjusted independently or relatively, then listen and check the full mix.
To normalize audio in DaVinci Resolve, select the audio clip or clips on the timeline, right-click and choose Normalize Audio Levels. Select the mode and target, choose Independent or Relative for multiple clips, and apply Normalize. Play the result in the full mix: matching peaks alone does not ensure that every voice sounds equally loud.
Blackmagic's Resolve 18 Beginner's Guide demonstrates this workflow on dialogue clips in the Edit page. Labels and available modes may vary with your Resolve version.
Normalize clips step by step
- Select the intended clips. Open the timeline and select the audio clips you want to adjust. Check that you have not included music or effects if you mean to adjust dialogue alone.
- Open the normalization dialog. Right-click one of the selected audio clips and choose Normalize Audio Levels.
- Choose a mode and target. For a sample-peak starting point, Blackmagic's training example uses Sample Peak Program and a target of −9 dBFS. This is an example for that lesson, not a universal delivery requirement. Confirm the settings offered by your version.
- Choose how selected clips are treated. Independent adjusts selected clips separately; Relative uses a common adjustment based on the selection. The former can help with clips recorded at different levels, while the latter preserves their existing relative level differences.
- Apply and review. Choose Normalize, play quiet and loud sections, then listen to transitions with music and effects active. Adjust individual clips or automation where a word or phrase still disappears.
Blackmagic's earlier normalization lesson explains that the relative and independent choices use the highest peak of the selection or individual clip peaks respectively. Save a working copy before making broad changes so you can compare your mix.
Independent or Relative: which should you use?
| Situation | Starting choice | What to listen for |
|---|---|---|
| Separate interviews recorded at different levels | Independent can give each clip its own peak adjustment. | A loud transient can still make one voice sound quieter overall; compare complete sentences. |
| A deliberate difference between clips | Relative preserves their existing level relationships through a common change. | Check that the loudest clip leaves room for the rest of the mix. |
| Voice with music and effects | Start with the relevant voice clips, not every track at once. | Listen in context and use clip gain or automation to keep words clear. |
Peak normalization versus loudness
Sample-peak normalization adjusts gain based on the highest sample in the selected audio. A single clap, plosive or click can determine the adjustment while the rest of the voice remains quiet. A peak target in dBFS is not the same as average or integrated loudness in LUFS. If your delivery specification uses a loudness target or true-peak limit, use the appropriate measurement and review the final mix against that specification.
Blackmagic's Resolve editor's guide describes peak and loudness normalization options. The peak versus loudness guide explains the distinction and why equal peaks can still sound different.
Fix common normalization problems
- One sentence is still too quiet: Check for a loud transient controlling the peak. Compare individual clips and use clip gain or automation for the quiet phrase; peak normalization alone does not level every word.
- Background hiss is more noticeable: Raising gain raises recorded noise along with speech. Improve the source or test appropriate cleanup on a copy before making up gain; check whether processing damages quiet words.
- The mix clips after normalizing: Check the combined output with all tracks and processing active. Individually safe clip peaks do not guarantee that overlapping tracks or later effects will stay below the output limit.
- Clips sound uneven despite matching peaks: Compare at the same listening level and adjust for perceived speech level. Consider the delivery specification and meter the complete mix, not just isolated clips.
- The recording is already distorted: Lowering or normalizing its level cannot restore clipped waveform detail. Return to a cleaner take where possible.
Prepare an audio file outside Resolve
The Saturalabs audio normalizer adjusts sample peak for a supported short MP3 or WAV in the browser. Download the result, import it into Resolve and check it in context. It is a file-preparation tool, not Resolve's timeline loudness meter or a guarantee of a compliant final mix.
Common questions
Does normalizing remove background noise?
No. A gain adjustment changes the level of the recorded signal, including its noise. Use source improvements or a suitable cleanup process when needed and compare the result with the original.
Should every clip be normalized to −9 dBFS?
No. Blackmagic uses that value in a specific peak-normalization training example. Choose levels for the material and delivery requirements, then check the complete mix.
Does normalizing change the original recording?
The described workflow adjusts selected timeline clips in your Resolve project. Keep your source files and a project copy, and check the final exported file separately.