Music ducking should sound like an arrangement decision, not a processor reacting to every syllable. Automatic side-chain compression can help with dense or repetitive material, but dramatic dialogue usually benefits from phrase-level automation and stem-aware control.
The aim is to reveal speech while preserving the cue's rhythm, tone and emotional continuity.
Start with Manual Phrase Rides
Identify the words that carry story information and lower music before they arrive, then recover after the phrase rather than between every word. Smooth ramps avoid audible holes beneath natural pauses.
Use music stems when available. Reducing a busy midrange element may create enough space without lowering bass, pulse or atmosphere that supports the scene.
- Automation anticipates important lines
- Recovery follows phrase and musical rhythm
- Stem balance is used before full-mix reduction
- Rides are checked at realistic playback level
Use Side-Chain Tools Selectively
If automation density becomes unmanageable, use gentle bus compression or dynamic EQ keyed from dialogue. Set attack and release to the cue, and limit the maximum reduction.
A broadband compressor changes the entire score, while a dynamic band can reduce only the masking region. Neither method understands dramatic importance, so keep manual control for key moments.
Dynamic EQ and music-carving starting points designed for dialogue-led audio post workflows.
View ResourceAvoid Pumping and Tonal Holes
Listen to music-only and full-mix passes. Repeated pumping may be hidden by dialogue but becomes obvious when the line stops or the scene moves to another language version.
Check M&E implications. Dialogue-keyed processing may not reproduce properly in a dialogue-free stem unless it is printed or documented according to the delivery design.
Practical QC Checklist
- Key words are clear without excessive music loss
- Gain recovery follows musical phrasing
- Side-chain reduction has a defined maximum
- Music-only continuity remains acceptable
- Stem and M&E behaviour is understood
Mixing perspective: Automation knows the story; a side chain only knows level. Use each for the job it can actually perform.
Combine broad manual rides, selective stem moves and restrained dynamic processing. The best ducking preserves the cue while making the dialogue feel naturally clear.