# How to add background music without drowning out your voice

Canonical URL: https://getvidiyo.app/blog/add-background-music-without-drowning-voice
Last updated: 2026-10-08
Author: Ajay Pawriya (https://getvidiyo.app/about)
Published: 2026-10-08
Category: Motion and audio

> TL;DR: Pick an instrumental track with little going on in the vocal range, set it well below the voice, and duck it further whenever someone speaks. Then measure instead of guessing: check the voice stays clearly above the music section by section, and bring the whole mix to about -14 LUFS for social video. Fade the music in and out.

Music makes a talking video feel finished. It also buries more voices than any other mistake, usually because the level was set on laptop speakers in a quiet room and never checked again. The fix is a few habits: choose a track that leaves room for speech, duck it, and measure the result rather than trusting your ears at one volume.

## How loud should background music be under a voice?

Quiet enough that every word stays clearly on top, everywhere in the video. Average loudness hides problems: a mix can measure fine overall while the chorus of the track swallows one sentence. Check the voice-to-music margin in short sections, and fix the sections that fail rather than lowering the whole track until it disappears.

These are the levels Vidiyo works with, as a reference:

| Setting | Value in Vidiyo | What it does |
|---|---|---|
| Starting music volume | About 0.2 (roughly 14 dB down) | A sensible starting point before ducking |
| Duck depth | 12 dB | How far music drops under speech |
| Duck timing | Down over 0.1 s before speech, up over 0.3 s after | Avoids pumping and late dips |
| Voice-over-music margin | Flagged below 8 LU | Any 3-second section where music competes |
| Mix target | -14 LUFS integrated, within about 1.5 LU | Overall loudness for social video |

[LUFS](https://getvidiyo.app/glossary/lufs) is a loudness unit that tracks how loud audio sounds to people rather than how high its peaks are. LU is the difference between two LUFS readings.

## How do I pick music that doesn't fight speech?

Pick instrumental tracks with steady energy and a sparse middle. Vocals in the music compete directly with your voice, and busy melodies in the same range as speech mask consonants. Pads, soft piano, light percussion and low-key beats sit under talking far more easily than a lead guitar or a sung hook.

- **No lyrics** under speech. Save vocal tracks for montage sections with no talking.
- **Steady, not dramatic.** Big builds and drops pull attention away from what is being said.
- **Match the length or loop it.** A track that ends mid-video, then restarts, is noticeable. Use a looping track or edit it to the cut.
- **Match the mood to the message,** not to what's popular this week.

## How do I duck music under a voice in any editor?

Put the music on its own track, set it low, then lower it further wherever someone speaks. Many editors have automatic ducking or a sidechain option; otherwise use volume keyframes at the start and end of each spoken section.

1. Place the music on a separate audio track under the dialogue.
2. Lower the music track until it sits comfortably under the voice in the busiest part of the song.
3. Turn on automatic ducking, or keyframe the volume down a few frames before each spoken section and back up after it.
4. Fade the music in at the start and out at the end.
5. Listen once on headphones and once on phone speakers. Phone speakers reveal masking that headphones hide.
6. Measure the integrated loudness of the export and adjust the master to your target.

Clean the voice before balancing. A noisy recording tempts you to push the music up to hide the noise, which buries the words. A high-pass filter, light compression and gentle noise reduction make the voice easier to keep on top.

## How do I add background music in Vidiyo?

Ask for it. "Put quiet lo-fi under my voice" is enough. The assistant picks a licence-cleared track by mood and length, spans the cut, loops it if the track is shorter than the video, adds fades, turns on ducking, then measures the mix and rebalances it.

Things you can type:

- "Add calm music under the whole video, keep it under my voice."
- "The music is too loud in the middle section."
- "Check the loudness before I export."
- "No music during the demo part."

The library is tagged by mood (upbeat, chill, cinematic, corporate, ambient and minimal) and says which tracks loop. It downloads the first time you use it. For a plain bed under speech, the assistant can also synthesize one: pads, drones and pulses with no melody to fight the voice.

Ducking follows the video clips with sound on your timeline, which is where filmed dialogue lives. Music set to duck drops 12 dB under them, easing down a tenth of a second before a clip starts and back up over three tenths of a second after it ends. A long silent stretch inside a clip stays ducked too. Cut those pauses and the music rises in the gaps. Narration imported as a separate audio file doesn't trigger ducking, but it's still measured as voice, so the balance step below keeps the music under it.

The measurement is what makes the difference. The assistant measures the mix as it will export: integrated loudness against -14 LUFS, sample peak, loudness range, and separate voice, music and other stems. It looks at the mix in three-second sections and flags music masking the voice, clipping, quiet dialogue, mismatched speakers and missing fades. Then it applies the recommended levels in one step. Each clip also has noise reduction, a six-band EQ, a de-esser and a compressor for cleaning the voice first.

## Should music play under the whole video?

Not always. Music under every second flattens a video, because nothing changes. Leaving it out for a key sentence makes that sentence land harder, and bringing it back on the next section signals a change of topic. Tutorials and screen demos often work better with music only in the intro and outro, so the viewer can follow the steps.

A simple structure works for most talking videos:

1. Music up for the first second or two, under the hook.
2. Ducked under the body of the video.
3. Out, or nearly out, for the one line you most want remembered.
4. Back up for the ending and the call to action, then a fade.

## What are the most common background music mistakes?

The common ones are easy to hear once you know them:

- **Music set by ear on headphones,** then too loud on phone speakers.
- **A track with vocals** under someone talking.
- **No fades,** so the music starts or stops abruptly.
- **The track's climax** landing on your most important sentence.
- **Music hiding a noisy voice** instead of cleaning the voice.
- **Video loudness way off target,** so it plays much quieter or louder than the videos around it.

Sound effects follow the same rule: they should sit under the voice, not on top of it. See [sound effects for short-form video](https://getvidiyo.app/blog/sound-effects-for-short-form-video) for levels and timing.

## Honest limits

- Vidiyo doesn't compose music; tracks come from its library, search or files you import.
- Ducking follows video clips with sound, so it also dips under their silent stretches. Narration imported as a separate audio file doesn't trigger it.
- Timelines max out at 5 minutes.

## Frequently asked questions

### How loud should background music be under talking?

Low enough that the voice is clearly on top in every section, not just on average. Vidiyo flags any three-second stretch where the voice is less than 8 LU above the music, and its starting volume for music is about 0.2, roughly 14 dB down, before ducking.

### What is audio ducking?

Ducking automatically lowers the music while someone is speaking and brings it back up in the gaps. It keeps the music audible in pauses without fighting the voice during speech.

### What loudness should I export social videos at?

Vidiyo measures the whole mix against -14 LUFS integrated, the target it uses for social video, with peaks kept below clipping. A mix within about 1.5 LU of the target plays as intended.

### Does Vidiyo duck music automatically?

Yes. Music set to duck drops by 12 dB under your video clips with sound, easing down just before each clip starts and back up just after it ends. The assistant can then measure the mix and rebalance it in one step.

### Is the music in Vidiyo safe to use?

The included tracks are licence-cleared, with source, author and licence kept for each one. Music you import yourself or find through search carries its own licence, so check it.

## Related

- [Music and sound effects that sit under your voice](https://getvidiyo.app/features/music-and-sound-effects)
- [Audio ducking](https://getvidiyo.app/glossary/audio-ducking)
- [LUFS](https://getvidiyo.app/glossary/lufs)
- [Sound effects for short-form video](https://getvidiyo.app/blog/sound-effects-for-short-form-video)
- [How to edit a talking-head video fast](https://getvidiyo.app/blog/edit-talking-head-video-fast)

Download Vidiyo for Mac (free during the beta): https://getvidiyo.app/download