Skip to main content
curious_Explorer7683
Participant
Nov 20, 2026 8:24 AM UTC
Answered

Best workflow for importing AI-generated audio stems and royalty-free BGM into Premiere Pro timeline?

Hi everyone,

I've been working on several video editing projects in Premiere Pro and trying to optimize my background music (BGM) workflow, especially to avoid YouTube copyright claims while keeping control over individual audio elements.

Recently, I've been experimenting with an AI audio workflow using Crea IA to generate royalty-free background tracks. What caught my attention is its stem separation and MIDI export features, which allow me to export separate stems (melody, drums, bass) and drag them directly into Premiere Pro's multi-track audio timeline. This makes cutting and side-chaining to dialogue much cleaner compared to standard pre-mixed MP3 tracks.

I'd love to ask the Premiere community:
1. What are your best practices or shortcuts in the Essential Sound panel when balancing AI-generated stems against dialogue?
2. Do you prefer dynamic linking with Adobe Audition for mastering these tracks, or do you keep everything directly on the Premiere timeline?

Looking forward to your insights!

CORCTON34033231
CORCTON34033231Correct answer
Community Expert
Nov 20, 2026 6:04 PM UTC

Hi, keeping the stems separate is exactly the right instinct — that's what makes balancing and ducking controllable. Here's the workflow that works well for this:

 

1. Import each stem as its own file (File > Import, or drag them into the Project panel) and put each stem on its own audio track — melody, drums, bass, and your dialogue on separate tracks. Rename the tracks so you can tell them apart at a glance.

 

2. Use the Essential Sound panel (Window > Essential Sound): select your dialogue clips and assign them the Dialogue type, then select the music clips and assign them Music. For the music, turn on Ducking and set it to duck against Dialogue — that automatically dips the BGM under speech, and you can adjust the sensitivity and how much it dips. That alone covers most BGM/dialogue balancing.

 

3. On Audition: I keep everyday balancing and ducking in Premiere, and only send a sequence to Audition (right-click the sequence > Edit in Adobe Audition) when I need deeper work — e.g. noise reduction, detailed multitrack mixing, or loudness work. Round-tripping every stem just for basic balance adds friction without much benefit.

 

One tip: keep a copy of the stems in a bin per project, and set your music track levels low first (-20 dB or so as a starting point), then let ducking do the work under dialogue rather than riding the faders by hand.

1 reply

CORCTON34033231
Community Expert
CORCTON34033231Community ExpertCorrect answer
Community Expert
Nov 20, 2026 6:04 PM UTC

Hi, keeping the stems separate is exactly the right instinct — that's what makes balancing and ducking controllable. Here's the workflow that works well for this:

 

1. Import each stem as its own file (File > Import, or drag them into the Project panel) and put each stem on its own audio track — melody, drums, bass, and your dialogue on separate tracks. Rename the tracks so you can tell them apart at a glance.

 

2. Use the Essential Sound panel (Window > Essential Sound): select your dialogue clips and assign them the Dialogue type, then select the music clips and assign them Music. For the music, turn on Ducking and set it to duck against Dialogue — that automatically dips the BGM under speech, and you can adjust the sensitivity and how much it dips. That alone covers most BGM/dialogue balancing.

 

3. On Audition: I keep everyday balancing and ducking in Premiere, and only send a sequence to Audition (right-click the sequence > Edit in Adobe Audition) when I need deeper work — e.g. noise reduction, detailed multitrack mixing, or loudness work. Round-tripping every stem just for basic balance adds friction without much benefit.

 

One tip: keep a copy of the stems in a bin per project, and set your music track levels low first (-20 dB or so as a starting point), then let ducking do the work under dialogue rather than riding the faders by hand.

Translate