How to Add AI Music to a Video (CapCut, Premiere Pro, DaVinci Resolve, iMovie)

Adding music to a video is a two-part job: getting a track that fits, and placing it so it supports the picture without drowning out anyone talking. Most guides only cover the second part and assume you already have a song. This one covers both: how to generate a track that is ready for editing, then exactly where to click in CapCut, Adobe Premiere Pro, DaVinci Resolve, iMovie and Canva to trim it, fade it and duck it under dialogue.

The generation steps use Song Creator Pro, a Windows app that generates music locally on your GPU. The editing steps work the same whatever generator you use.

Part 1: Generate a Track That Is Easy to Edit

A few decisions at generation time save a lot of fiddling in the editor.

Step 1: Decide the length

Check how long your video (or the section that needs music) runs, and generate a track a little longer than that. Trimming a track down is easy. Stretching one is not. In Song Creator Pro, set Duration in Music Details (anything from 10 seconds to 10 minutes on ACE-Step 1.5).

Song Creator Pro does not sync music to a video timeline or cut to an exact length for you, so plan on trimming in your editor. If you also set BPM, you can predict where clean cut points will fall: at 120 BPM a bar of 4/4 lasts exactly 2 seconds, so 4 bars is 8 seconds, 16 bars is 32 seconds.

Step 2: Go instrumental if anyone talks

If the video has narration, interviews or a voice-over, generate instrumental music. Sung lyrics and spoken words compete for the same attention, and viewers lose both. Leave the lyrics field empty or tick Instrumental, and write a style description that leaves room for a voice:

Warm cinematic underscore, soft piano and strings, gentle pulse,
steady dynamics, unobtrusive, background music for narration,
instrumental

Upbeat lo-fi hip hop, mellow keys, soft drums, relaxed and steady,
no lead melody, background for talking, instrumental

Words like "steady dynamics", "unobtrusive" and "no lead melody" steer the model away from big drops and busy hooks that pull focus from speech. Set Batch Size to 3 or 4 and pick the one that sits best under your voice.

For a montage with no dialogue, the opposite applies: a track with a clear build and a strong ending can carry the whole edit. Vocals are fine there too. See the prompting guide for more on style descriptions, and the prompt library for ready-made ones.

Step 3: Use stems if one part fights the voice

Sometimes you get a track you love except for a lead synth or guitar line that sits right in the vocal frequency range. Instead of regenerating, run it through Song Creator Pro's stem separation (Demucs, with four separation models). You can drop the lead and keep drums, bass and pads under the talking sections, then bring the full mix back for the montage at the end. Our guide to extracting vocals and instruments walks through it.

Step 4: Export as WAV

Song Creator Pro exports MP3, WAV and FLAC. Use WAV for editing. The video export will compress the audio once at the end anyway, and starting from a lossless file avoids stacking two rounds of compression.

Part 2: The Editing Basics (Any Editor)

Whatever software you use, the job is the same four moves:

1. Import the WAV and put it on its own audio track, under the video.
2. Trim the start so the music lands where you want it, and the end so it finishes on a phrase rather than mid-note.
3. Fade out the last 1 to 3 seconds (and fade in if the track starts abruptly).
4. Duck the music under any speech, typically 12 to 20 dB below the voice, rising in pauses.

Here is how each editor does it.

CapCut (Desktop)

Import and place

1. Click Import in the media area at top left, or open the Audio tab and choose your file.
2. Drag the track onto the timeline below your video clips.

Trim

• Drag either edge of the audio clip to shorten it, or move the playhead and press Ctrl+B (Split) to cut, then delete the leftover piece.

Fade

• Select the audio clip. In the panel on the right, under Basic, set Fade in and Fade out durations. (On some builds these sit under an Audio tab in the same panel.)

Duck under voice

• CapCut desktop does not offer a reliable one-click ducking control that we could confirm. Do it with keyframes: select the music clip, move the playhead to just before speech starts, and click the keyframe diamond next to Volume in the Basic panel. Move a little later, lower the volume and add another keyframe. Repeat where speech ends to bring it back up.
• For a video that is talking all the way through, simply lower the whole music clip's Volume until the voice is clear.

CapCut mobile: tap the audio clip, then use Volume and Fade. Recent mobile versions also offer an automatic ducking option for speech; if yours does, it is the quickest route.

Adobe Premiere Pro

Import and place

1. Drag the WAV into the Project panel (or use File > Import).
2. Drag it from the Project panel to an empty audio track on the timeline.

Trim

• Drag the clip edges, or press C for the Razor tool, click to cut, then select and delete the extra.

Fade

• Move the playhead to the end of the clip, select it and press Ctrl+Shift+D (Cmd+Shift+D on Mac) to apply the default audio transition, Constant Power. You can also drag Constant Power from Effects > Audio Transitions > Crossfade onto either end of the clip and drag its edge to change the length.

Duck under voice (automatic)

1. Open Window > Essential Sound.
2. Select your voice clips and click Dialogue.
3. Select the music clip and click Music.
4. Tick Ducking, set Duck against to the Dialogue icon, and adjust Sensitivity, Reduce by (start around 12 to 18 dB) and Fades.
5. Click Generate Keyframes. Premiere draws volume keyframes on the music that dip under every line.

Two gotchas: enabling Ducking replaces any volume keyframes you drew by hand on that clip, and some users have reported Generate Keyframes failing when the dialogue track is soloed, so unsolo before you click. Recent versions also expose clip audio controls in the Properties panel, but Essential Sound is still where the ducking settings live.

DaVinci Resolve

Import and place

1. On the Media or Edit page, drag the WAV into the Media Pool.
2. Drag it onto an audio track in the Edit page timeline.

Trim

• Drag the clip edges, or press B for the Blade tool, cut, and delete.

Fade

• Hover over the audio clip on the Edit or Fairlight page. White fade handles appear at the top corners. Drag them inward to create a fade-in or fade-out.

Duck under voice: the Ducker effect Resolve 19 introduced a dedicated Ducker track effect, and it is in the free version. Add it to your music track (from the Fairlight FX in the Effects Library or the track's effects), choose your dialogue track as the source, then set how much to reduce and how quickly it recovers. This is the easiest route for most people.

Duck under voice: the classic sidechain method On the Fairlight page:

1. On the dialogue track's mixer strip, double-click Dynamics and set the compressor's sidechain to Send.
2. On the music track's Dynamics, enable Compressor and set it to Listen, pointing at the dialogue.
3. Adjust threshold and ratio until the music dips clearly when someone talks.

If you have several dialogue tracks, route them to a bus and sidechain from the bus. Resolve Studio's AI audio tools can also handle ducking as part of an automatic mix.

iMovie (Mac)

Import and place

1. Click Import Media or drag the WAV into your project's media.
2. Drag it into the background music well (the area under the timeline marked with a music note) so it runs under the whole movie, or drop it directly under a clip to attach it there.

Trim

• Drag the clip's edges, or select it and use Modify > Split Clip at the playhead.

Fade

• Hover over the audio clip. Drag the small fade handles at the start and end inward.

Duck under voice

1. Select the clip that should stay loud (your voice-over or the video clip with speech).
2. Click the Volume button above the viewer.
3. Tick Lower volume of other clips and drag the slider to set how much the music drops.

iMovie on iPhone and iPad has no equivalent checkbox. Lower the music clip's volume manually and use the fade controls.

Canva (Brief)

Canva's video editor is fine for short social clips.

• Add: open Uploads in the left panel, upload the WAV, and click it to add it to the timeline (Canva's own library is under Audio).
• Trim: select the audio track on the timeline and drag its handles.
• Fade: select the audio track, click Fade (audio effects) in the toolbar above the editor, and drag the Fade in and Fade out sliders.
• Duck: there is no automatic ducking that we could find. Lower the music track's Volume from the toolbar, or split the track and lower only the parts under speech.

Quick Reference

Editor	Fade	Automatic ducking
CapCut desktop	Basic panel: Fade in / Fade out	No (use Volume keyframes)
CapCut mobile	Fade	Auto ducking on recent versions
Premiere Pro	Ctrl+Shift+D, Constant Power	Essential Sound: Ducking > Generate Keyframes
DaVinci Resolve	Clip fade handles	Ducker effect, or sidechain compressor on Fairlight
iMovie (Mac)	Clip fade handles	Volume > Lower volume of other clips
Canva	Fade (audio effects)	No (manual volume)

Mixing Tips That Make a Difference

• Check on a phone. Most of your audience will hear this through phone speakers or earbuds. Bass-heavy music that sounds fine on headphones can bury a voice on a phone.
• Fade between scenes, not on beats. If you cut the music between two scenes, cross the cut on a downbeat and use a short crossfade of a few frames.
• Let the music breathe. Bring the music up to full level in gaps longer than a second or two. A video where the music never rises feels flat.
• Do not normalise the music to the voice. Background music should sit well below the dialogue; aim for the voice to carry the loudness of the video.

If you need short pieces rather than a full bed, such as a channel intro or a logo sting, see how to make intro and outro music with AI. And if you are worried about copyright claims after uploading, read will AI music get a Content ID claim and our guide to AI music for YouTube Shorts and TikTok.

---

Ready to score your next video? Get Song Creator Pro on the Microsoft Store for $49.99 (free trial available) or on itch.io for $44.99. One-time purchase, unlimited generations, WAV export and stem separation, with a commercial licence for your videos included.