Separate overlapping speakers

Last updated on Nov 10, 2026

Learn how to separate overlapping speakers in Adobe Premiere using AI, so you can edit each speaker's audio independently.

Premiere allows you to identify and separate overlapping speech in an audio clip. The selected clip is analyzed to generate separate audio tracks, allowing you to adjust each speaker independently while preserving the original audio.

Note

Separate Crosstalk is a premium feature that uses one generative credit per second of processed audio. Trim your clip to the overlapping speech section to reduce credit usage.

Separate Crosstalk automatically adds a two-second audio handle to each side of the selected clip. This additional processing time is reflected in the cost displayed before processing. To adjust the audio handle duration between 0 and 5 seconds, go to Preferences > AI Models > Separate Crosstalk > Audio Handles.

Select the audio clip in the Timeline.

Trim the audio clip to isolate the section containing overlapping speech.

Each separation supports audio clips between 1 second and 10 minutes in duration. If your clip exceeds 10 minutes, trim it before proceeding.

Right-click the selected audio clip or segment, and then select Separate Crosstalk.

Alternatively, you can select Separate Crosstalk in the Properties panel or go to Clip > Separate Crosstalk from the menu.

The Clip menu is open and the Separate Crosstalk option is selected and the same option is selected in the Properties panel.
Access Separate Crosstalk from the clip context menu.

A. Separate Crosstalk in the clip context menu B. Separate Crosstalk in the Properties panel 

Select Separate in the Separate Crosstalk dialog box.

The Separate Crosstalk dialog displays a description of voice separation, credit consumption, and a Separate button.
Review the processing details before separating the audio.

Review the generated tracks for each detected speaker and the background audio.

The Premiere timeline displays the original audio track (A) and three separated green audio clips (B, C, and D) beneath it.
Separate tracks for each speaker and the background audio.

A. Disabled original track B. Speaker 1 C. Speaker 2 D. Background track 

The original audio track is automatically disabled. Review and edit the generated speaker and background tracks as needed.

Note
  • Separate Crosstalk doesn't have a default keyboard shortcut. You can assign one in the Keyboard Shortcuts dialog box.
  • You can separate speakers for one audio clip at a time. Multi-clip selection is not currently supported.