The couple's speech, the priest, an interview — with music underneath. Voice Duck finds where people actually talk

And lowers the music exactly there, with real volume keyframes you can still adjust by hand.

How to use

1. Open the sequence. Window → Extensions → DAEDAL Voice Duck.

2. Tick the VOICE tracks (lav, recorder, mic) and the MUSIC tracks. The panel guesses them from the clip names.

3. "Analyze voice": the drawing shows where speech was found (gold) and how the music will move (blue line).

4. Adjust if needed: Lower the music by — 12 dB is a good start for speeches; 15–18 dB for quiet voices. Sensitivity — medium for most cases; low if the music from the speakers leaks into the lav mic; high only for soft, distant speech. Duck ahead by / Come back in — how fast the music moves (attack / release). Don't come back in pauses shorter than — keeps the music down between sentences.

5. "Apply ducking to music" writes volume keyframes on the music clips (existing ones are replaced if ticked). "Remove ducking" deletes them and restores the previous music level. Ctrl/Cmd+Z works too.

Analysis runs inside the panel — nothing is uploaded anywhere.

AUDIO INSIDE VIDEO FILES needs ffmpeg. It is included in DAEDAL Sync (free) — install it and the panel finds it.

WAV/MP3 files from recorders work without it.

Install

Mac: quit Premiere, unzip, right-click "Instaleaza (Mac).command" → Open → Open.

Windows: quit Premiere, unzip, double-click "Instaleaza (Windows).bat".

Then Premiere → Window → Extensions → DAEDAL Voice Duck.

Updates

When the panel opens it checks daedal.art for a newer version. If there is one, a bar shows what changed and an "Update" button: one click downloads it from DAEDAL's GitHub, checks it and reloads the panel.

Offline, the panel works normally.

License: free for any use. Provided as is.