Enhance Audio overview

Enhance Audio separates production audio into individual layers for dialogue, music, ambience, and sound effects. You can enhance dialogue, reduce reverb, and rebalance the other elements in Premiere. It combines on-device source separation with Enhance Speech, giving you more control than a single blended noise-reduction pass.

What can you adjust?

Enhance Audio provides controls for each separated audio layer.

Dialogue

You can adjust the dialogue using the following controls:

• Volume: Adjust the level of the dialogue independently.
• Enhancement: Control the amount of dialogue enhancement. At the lowest setting, the dialogue retains the original sound. Increase the setting to apply more enhancement.
• Reverb: Reduce reverb from the dialogue. The amount of change depends on the source recording and affects only the reverb on the speech.

Music, ambience, and sound effects

Music, ambience, and sound effects each have a Volume control. You can increase or decrease the level of a layer or mute it completely.

The clip waveform updates as you adjust the layers, reflecting your changes in the timeline.

Enhance Speech models

Enhance Audio supports two Enhance Speech models:

• Enhance Speech V1: Runs fully on-device and includes source separation. It does not use generative credits.
• Enhance Speech V2: Uses the highest-quality voice model and runs in the cloud as part of Firefly. It uses two generative credits per minute for speech enhancement.

Choose a default model in Preferences, or switch between Enhance Speech V1 and Enhance Speech V2 in the Enhance Audio settings in the Properties panel.

Source separation runs on-device. Generative credits are used only when Enhance Speech V2 is used for speech enhancement.

Edit enhanced clips

Enhance Audio includes five seconds of handles on either side of the analyzed clip. You can trim or extend an enhanced clip within these handles without triggering reprocessing.

If you extend the clip beyond the available handles, Premiere processes only the newly added material instead of reprocessing the entire clip.

When you use Enhance Speech V2, processing newly added material uses generative credits at the rate of two credits per minute for the speech enhancement portion.

Split enhanced audio into separate clips

You can split an enhanced result into separate clips, with speech, music, ambience, and sound effects placed on the timeline as individual clips.

Once separated, you can edit, move, and adjust each element independently. This is useful when you need to edit individual audio stems or prepare separate audio elements for another editor or sound mixer.