# Add background music to a video and keep the voices

> Markdown version of https://stella.sh/how/video/add-music-to-video for agents and LLMs.
> Put a song under your video quietly enough that voices stay clear, and fade it out at the end: iMovie, Clipchamp, ffmpeg, or Stella (stella.sh) on your own computer.

Put the song on its own track under the video, turn it well down, trim it to the video's end and add a fade. iMovie's background music well does it on a Mac, and Clipchamp on Windows. The usual mistakes are music that drowns the voices, replaces them, or runs on past the end. Stella (stella.sh) mixed a 2½-minute song under a 40-second beach video, about 14 dB under the voices, and faded it out over the last 5 seconds.

- Checked: 14 November 2026
- Stella does it on: Windows, Mac, Linux

## Stella doing it

Recorded 14 November 2026 on Linux (Omarchy) with Claude Code with Opus, on a Claude plan.
Result: The new video in under a minute. Checked afterwards against an answer key written before the run: 40 seconds (not stretched to the song), the picture unchanged, the song in the soundtrack, the voices kept and 14 dB louder than the music, the music faded to 12% of its level by the last second, and the original video and song unchanged. The beach video has two synthetic voices over wave sounds, and the song is a synthesised instrumental by a made-up band; both were made for this run. The work is sped up 3×. Nobody listened to the mix; it was measured.

Video (46 s): https://stella.sh/media/proofs/video/add-music-to-video/proof.mp4

Transcript:

- 0:00 Stella on a Linux computer. A beach video is in Videos and the song Sunny Days in Music.
- 0:01 The prompt: add the song quietly under the talking and fade it out at the end.
- 0:14 Stella replies the song will sit under the voices and fade out, saved as a new file.
- 0:20 Measuring the voices and the song, then mixing.
- 0:31 Done: the song under the voices, fading over the last 5 seconds; levels measured, not listened to.

What Stella does:

1. Finds “Sunny Days” in Music (2:30) and “Beach day” in Videos (40 seconds), the only match for each
2. Measures how loud the voices and the song are
3. Mixes the song from its start under the original sound at a low level, trimmed to the video's length and fading out over the last 5 seconds
4. Copies the picture across untouched and saves Beach day (with music).mp4 next to the original
5. Measures the mix and says the levels were checked by measurement, not by listening

## The prompt to paste

```text
Add the song Sunny Days from my Music folder as background music to the beach video in my Videos. Keep the sound of us talking, with the music quieter underneath, and fade the music out at the end.
```

## Add the music yourself

1. **Mac: iMovie.** Drag the video into a new project, then drag the song into the background music well below the clips. Click the song, use the Volume button to bring it well down, and drag the small fade handle at its end leftwards. Then File, Share, File.
2. **Windows: Clipchamp.** Drag the video and the song onto the timeline. Select the song, lower its volume in the Audio tab (try 20–30%), set a fade-out in the Fade tab, and drag the song's end to line up with the video's end. Export.
3. **ffmpeg.** `ffmpeg -i video.mp4 -i song.mp3 -filter_complex "[1:a]volume=0.08,afade=t=out:st=35:d=5[m];[0:a][m]amix=inputs=2:duration=first:normalize=0[a]" -map 0:v -map "[a]" -c:v copy -c:a aac out.mp4`. Set `st` to the video's length minus the fade.
4. **Set the level by ear.** Play the busiest talking part. If you have to strain to hear words, the music is too loud. Speech usually needs to sit 10–15 dB above background music, which is quieter than most people first guess.
5. **Use music you can share.** Instagram, TikTok and YouTube may mute or block videos with copyrighted songs. Their own music libraries, or royalty-free tracks, avoid that.

## Where Stella stops

- The music level was set by measurement. Stella said it hadn't listened and offered to make the music louder or softer.
- The song started from its beginning; picking a different section of the song, or ducking it further while people talk, would be a follow-up request.
- File names and measurements of the sound went to the AI model you use in Stella. The video and song were mixed on your computer.
- Recorded on Linux with ffmpeg already installed. On a Mac or Windows PC Stella would usually download ffmpeg first, which isn't shown.

## Questions

**How do I add music without losing the original sound?**
Mix the song in as a second track under the existing sound rather than replacing it. iMovie, Clipchamp and Stella all keep the original voices.

**How loud should background music be?**
Well under the voices: about 10–15 dB quieter. In this run the voices came out 14 dB over the music.

**What if the song is shorter than the video?**
Loop it, or add a second song. Ask Stella to loop it with a crossfade.

**Can I make a slideshow with music instead?**
Yes: see [make a slideshow with music](https://stella.sh/how/video/make-slideshow-with-music).

**Is Stella free for this?**
Yes. Free with the ChatGPT or Claude plan you already have. Sign in with your ChatGPT plan, or use your Claude plan through Claude Code, and this task costs nothing extra. New accounts also get a small free trial of Stella’s own models.

**Can I control what Stella changes on my computer?**
Yes. Stella works on Windows, Mac and Linux with your own files and shows each step in the chat, so you can watch it and stop it at any point. Tell it what it may not do, as the prompts on these pages do (“don't delete anything without asking”). In the Downloads recording it deleted nothing until it was told to go ahead. It also refuses commands that would wipe a whole disk, your home folder or system folders.

Sources: [Apple: Add music and sound clips in iMovie on Mac](https://support.apple.com/guide/imovie/mov91a895a64/mac), [Microsoft: How to add audio or music to videos (Clipchamp)](https://support.microsoft.com/en-us/clipchamp/how-to-add-audio-or-music-to-videos), [FFmpeg filters: amix and afade](https://ffmpeg.org/ffmpeg-filters.html#amix).

## Related problems

- [How to make a slideshow with music from your own photos](https://stella.sh/how/video/make-slideshow-with-music.md)
- [Increase the volume of a quiet video without distortion](https://stella.sh/how/video/make-video-louder.md)
- [Make all your songs the same volume](https://stella.sh/how/audio/normalize-audio-volume.md)
- [Add burned-in subtitles to a video for TikTok](https://stella.sh/how/video/burn-subtitles-into-video-for-tiktok.md)
- [Cut silences out of a video automatically](https://stella.sh/how/video/cut-silences-from-video.md)
- [Compress a video for Discord](https://stella.sh/how/video/compress-video-for-discord.md)

Get Stella: https://stella.sh/download. Free with the ChatGPT or Claude plan you already have.
