Audio Ducking for Narrated Videos: A 6-Step Workflow

What Is Audio Ducking?

A viewer will forgive an average shot. They won't sit through narration they can't make out. Subtle music is fine, but people drift once the voiceover gets hard to follow.

Audio ducking is how you keep both. It lowers the volume of one track whenever another track is active, then lifts it back when that track stops. In a narrated video, the music is the ducked track and the voice is the trigger.

What audio ducking does

Ducking reacts to the voice itself. That makes it different from a fixed fade, and it scales well across long recordings. Under the hood, most editors use sidechain compression and simply label it "auto-ducking".

The result: the music keeps the mood, and the voice keeps the viewer.

The 6-Step Audio Ducking Workflow

Turning on auto-ducking is one click. Getting it to sound right takes these six steps.

1. Identify when music masks the voice

Speech clarity lives mostly between 1 and 4 kHz, with the peak around 2 kHz. Consonants like k, p, s and t sit in the 2 to 4 kHz range. When music is busy in that same band, words blur even at a normal volume.

Play each scene with the music at full level under the narration first. Note every line you have to rewind to catch. Those are your masking spots.

2. Pick music that leaves room for the voice

A track with lyrics will always fight a voiceover. Lead instruments like flute, sax, solo violin and screaming guitar also tend to cover speech. Sparse, instrumental tracks need less ducking than busy ones.

Still clashing? A 5 to 10 dB EQ cut in the music between 1 and 4 kHz improves intelligibility.

3. Plan where the music should dip and return

For each narration block, set this keyframe pattern on the music track:

1. A keyframe shortly before the narration begins.
2. A keyframe where the music reaches its lowered level.
3. A keyframe near the end of the narration, still at the lowered level.
4. A keyframe after the last word, where the music is back up.

Start the fade slightly before the line so the first syllable isn't buried. Bring the music back after the final word, not during the tail of the sentence.

Keep the music low through short pauses between connected lines. Let it rise in meaningful breaks. Treat the intro and outro on their own: the music can carry the opening, then step down as the voice comes in.

4. Avoid pumping and abrupt level changes

With manual keyframes, pumping usually comes from raising and lowering the music in every brief pause. Abrupt changes come from fades that are too short. Music should fade down and come back gradually, not jump.

Listen to the first few words of each line. If the music stays loud too long, shorten the fade-down. If it seems to vanish, make the fade longer.

There's no universal ducking depth. A 6 to 12 dB reduction is a common starting range. In one Premiere Pro tutorial, lowering the music to -18.5 dB between keyframes was enough to hear the voiceover clearly. When in doubt, pick the quieter music.

5. Check the mix on headphones and speakers

Headphones can make a mix feel fuller and wider than it sounds on speakers or phones. Use headphones to hear detail, then speakers to judge balance and clarity. Collapsing the mix to mono helps you hear what phone playback will do.

Run one phone or laptop check before export. Focus on the first narration line, dense music sections, quiet passages and the outro. If you strain to hear a word there, pull that keyframe lower.

6. Review the final scene-to-scene audio balance

Volume moves can sound clean scene by scene and still jump in energy across the full video. Watch the whole piece once at a normal listening level and mark every jump.

Level the narration before you set keyframes, so the voice doesn't swing between quiet and loud sections. Then reuse one music level and fade pattern across similar scenes as your starting point. Adjust only where a scene needs it.

Keep Voice and Music in One Place