How do I auto duck music under a voiceover?

Upload the video and say which layer stays in front. Sparki follows the voiceover's timing, dips the music for the length of each spoken section and lifts it again between them. Ask for a deeper or gentler dip on any section and that range alone changes.

Will ducking make the music inaudible under speech?

No. The point is headroom, not silence — the track stays present under the voice at a level that keeps sentences effortless to follow, then returns to full energy in the gaps. If a section needs the music almost gone, say so and that section drops lower.

Is audio ducking the same as noise reduction?

No. Ducking only manages the volume relationship between layers that already exist in your mix. It does not remove hiss, wind or background noise, separate a voice out of a finished track, or repair distorted audio — those are different jobs, and Sparki will not pretend ducking did them.

Does ducking work if my video has no music yet?

Ducking needs two layers to balance. If the video has speech but no track, ask for background music first and Sparki adds one, then ducks it under the voice in the same pass — the levels are set from the start instead of fixed afterwards.

Do I need to set the ducking level myself?

No. Sparki picks a dip depth that keeps speech clear and returns the bed between lines, then normalises the export so the balance survives phone speakers. You steer with words — deeper, gentler, more music during the outro — rather than by typing decibel values.

Audio Ducking

Mixes almost always fail the same way: music laid at full level, speech fighting to be heard, and no hours left to draw volume keyframes by hand. Ask for ducking and Sparki rides the music level against the voice — down while someone speaks, up between lines, gone cleanly at the end.

What Ducking Does To Your Mix

The volume relationship between voice and music, handled for you.

The Bed Drops When The Voice Starts

Sparki finds the speech — recorded narration, a generated voiceover, someone talking on camera — and lowers the music beneath it for exactly as long as the sentence runs. No keyframes to draw, no decibel guessing: the level follows the voice, and a phrase that runs long keeps its headroom the whole way.

It Returns In The Gaps, Fades At The End

Between sentences the bed breathes again, so the edit keeps its energy instead of flattening. When the speech stops for good, the music comes up for the outro or fades out where you say. The result sounds like someone rode the fader all the way through — someone did, just not you.

Balanced Across Every Layer

Ducking is one part of the mix. Sparki also keeps effects, ambience and original audio in their lane — speech on top, music beneath, everything else out of the way — and normalises the export so the balance survives a phone speaker, not just the machine that made it.

Where Auto Duck Music Matters

Six mixes that go wrong without it.

Voiceover Over Music

Scripted narration, full bed

Explainers, ads and faceless-channel videos where a narration track plays over continuous music. Sparki dips the bed under every narration section and restores it between them, so the read stays forward without losing the momentum the music provides. Any video where the words are the content and the music is the mood.

Commentary Over Game Audio

Voice versus loud capture

Gaming highlights and stream recuts where the commentator talks over loud in-game sound. Sparki holds the commentary on top and pulls the game audio down while the voice runs, letting the play's own sound back in during the gaps. Clips where both the reaction and the action need to be heard.

Instructions Over A Music Bed

Recipe and workout step audio

Cooking, fitness and DIY videos where spoken steps sit on top of a track chosen for energy. Sparki keeps the bed low through each instruction and lifts it in the quiet stretches between steps, so the pace never dies mid-recipe. Step-by-step content filmed to music.

Interviews With A Music Bed

Long answers, steady level

Cutdown interviews and testimonial edits with music underneath the conversation. Sparki rides the bed under long answers without pumping, and stays out of the way when the room tone carries a pause. Testimonials and case-study videos with underscore.

Multi-Voice Clips

Two speakers, one track

Podcast clips, duets and dialogue edits where several voices share one music bed. Sparki follows whoever is speaking and keeps the bed clear of both voices, so handoffs between speakers never leave the music louder than the next person. Conversation formats cut for social.

Outros & End Cards

Music rises after the speech

The final seconds of promos and episodes, where the call to action ends and the music should close the video. Sparki lets the bed swell once the speech finishes and fades it on the end card, instead of chopping the track off mid-phrase. Endings that feel finished rather than truncated.

Who Needs Audio Ducking

Four people whose voice keeps losing to the music.

Podcast clipper

Full-energy theme under talk segments

Cuts podcast moments for social with the show's theme track underneath. Sparki ducks the theme under every spoken line and brings it back between clips, so the hook is audible in a feed where most viewers never turn the sound up twice.

Stream recutter

Voice over loud game capture

Pulls highlights from VODs where the commentator drowns in their own game audio. Sparki drops the capture under the reaction, restores it for the play and exports a clip where both the call and the clutch survive.

Recipe video creator

Steps spoken over a backing track

Films to music because the kitchen audio is unusable. Sparki keeps the track beneath every instruction and lifts it while the oven does the work, which is what separates a followable recipe from a merely pretty one.

Agency social manager

Client reels with voice and track

Delivers dozens of reels a month where a spoken offer sits over licensed music. Sparki applies the same voice-forward balance to every client cut, so nothing ships with the offer buried and no afternoon is lost drawing volume curves.

How To Auto Duck Music In A Video In 3 Steps

Upload, say what stays on top, export the balanced mix.

1
Upload The Video
Send the file or paste a link. Videos with narration, on-camera speech, a generated voiceover or a baked-in track all work — Sparki sorts the layers before touching levels.
2
Say What Should Stay On Top
Name the layer the viewer must not miss — the voiceover, the commentary, the instructions. Sparki lowers the music under that layer's speech and brings it back in the gaps, with fades where the track starts or ends.
3
Listen Back And Export
Play the mix, ask for a deeper dip under a specific section or more music in the quiet parts, then export with the balance normalised for platform playback.

Explore More Sparki Tools

Ducking is one mix control. Music selection, sound effects and captions share the same upload.

Audio Ducking FAQ

It lowers one audio layer — usually the music — whenever another layer, usually speech, is playing, and restores it in the gaps. Sparki automates the whole relationship: the bed drops under every sentence, breathes between lines and fades cleanly at the end, with no volume keyframes drawn by hand.

Let The Voice Lead, Let The Music Follow

Upload the video and name the layer that matters most — Sparki ducks the music under it, restores it between lines and exports a mix that holds up everywhere.

Duck your music automatically