What did AI actually solve?

Enhance Audio makes a mixed recording editable by sound class. Try it on a copy of one clip, choose the appropriate processing option and keep only changes that sound right.

01

What changed

A single recording can contain a voice, music and the room around it. Adobe’s earlier Enhance Speech workflow offered a Mix Amount slider to blend the cleaned-up and original audio. The Enhance Audio public beta, announced September 2, 2026, adds separate controls for dialogue, music, ambience and sound effects.

The sound classes can also become individual clips in the timeline. That gives an editor a way to work on layers extracted from one recording, alongside the separate-track mixing Premiere already supported. Adobe’s September 8 product update corroborates the new independent controls.

02

What this could change for you

For an interview or a personal video, you could try lowering the extracted music while leaving the dialogue level alone, or adjust the atmosphere separately. You choose the balance after separation; the software does not make every mixing decision for you.

The documented release is Premiere Beta 27.0, build 36. Install the beta offered to your account through Creative Cloud’s Apps → Beta section and check computer compatibility and subscription access. Adobe says separation runs on-device. With Enhance Speech V1, the workflow uses no generative credits; optional V2 voice enhancement runs in the cloud and uses two generative credits per minute.

Try it yourself

Try separating one mixed recording

A compatible computer and Adobe account with access to Premiere desktop beta containing Enhance Audio. Check the current subscription and beta availability. Choose local V1 or optional cloud V2 before processing; V2 requires cloud access and generative credits.

  1. Install or open the available Premiere beta through Creative Cloud Apps → Beta. Work on a copy of a short recording you have permission to process.
  2. Select the audio clip, open Properties → Audio and inspect the Enhance settings. Choose the local V1 or cloud V2 voice option, then enhance the clip.
  3. Listen to each class while adjusting its volume. If useful, split the enhanced result into separate timeline clips; keep the original and compare before retaining changes.

What to watch for: Suggested from official instructions, not firsthand tested. Sound classes do not imply perfect isolation or separate tracks for every speaker or instrument. Check the beta’s current controls and V2 credit conditions.

Read Adobe’s Enhance Audio beta workflow ↗

Suggested activity. Official guidance checked .

03

What it does not prove

These are broad sound classes, not a promise of a separate track for every person or instrument. Keep the original recording and listen to the separated layers before relying on an edit. The beta is still changing.

The evidence is Adobe’s release and access documentation. We have not tested the feature firsthand or independently measured separation quality, word fidelity or time saved. The local V1 option and cloud V2 option have different processing and credit conditions.

Primary sources

Now in Beta: Enhance Audio in Premiere

Adobe Premiere product-team announcement · September 2, 2026

View the announcement ↗