Put music under a voice, without uploading either
Load the recording, lay a music track beneath it, and let the music step out of the way whenever someone speaks. Both files stay on your machine, which matters when the episode is not out yet.
Music beds usually mean uploading both the recording and the track to someone else. Here the mix is made on your machine, so a rough cut of an unreleased episode never leaves it.
How to add background music to a recording
- Open the editor and drop your voice recording in first. It becomes the main track, the one the music will duck under.
- In the Music overlay panel, load the music. It is drawn in purple underneath the waveform, so you can see where it sits against the speech.
- Choose where it starts: at the very beginning, or at the playhead if you want the voice to open alone and the bed to come in later.
- Set the volume. It starts at -14 dB, which is roughly the level a bed sits at under speech, then adjust by ear.
- Turn on loop if the music is shorter than the recording, and set a fade in and fade out in seconds so it does not start or stop abruptly.
- Press preview the mix and listen before you commit to anything. When it sounds right, mix it in and export as MP3 or WAV.
What ducking does, and why it is on by default
Ducking is the thing that separates a music bed from a mess. Without it you pick one volume: loud enough to hear in the gaps means it fights the voice, quiet enough to stay out of the way means it vanishes. Ducking solves that by changing the music volume continuously, following the speech.
The editor watches the loudness of your voice track and pushes the music down while you are talking, then lets it come back up when you stop. By default it dips the music 10 dB under speech, and you can set anything from 0 to 30 dB.
Two details make it sound like a human did it rather than a gate. It measures the voice over a short window rather than sample by sample, so an ordinary consonant does not trigger it. And it ducks quickly but recovers slowly, which is how broadcast ducking has always worked: the music gets out of the way the instant you speak, then eases back over the pause rather than leaping up between two sentences.
Picking a level that works
- Start at the default and listen on something ordinary. Beds that sound perfect in good headphones often disappear on a phone speaker, which is where most people will hear it.
- If the music is still distracting, duck deeper before you turn the whole bed down. A deeper duck keeps the music present in the gaps, which is the part that is doing the work.
- Music with a busy middle fights a voice no matter what you do with the fader, because that is where speech lives. A sparse track at a modest level beats a dense one pushed quiet.
- Level the voice first. Ducking follows the voice, so if the speech wanders up and down, the music will wander with it. Even out the recording before you add the bed.
What this will not do
- It is not a multitrack studio. One main recording and one overlay at a time. For a voice, a bed and three sound effects on separate tracks, use a desktop editor.
- It will not find you music. There is no built-in library, so bring your own file.
- It cannot give you the rights to a song. Using commercial music in something you publish needs a licence, whatever tool made the mix. That is between you and the rights holder, and no audio editor changes it.
Common questions
How do I put background music under a voiceover?
Load the voice recording as the main track, then load the music in the Music overlay panel. Set where it starts and how loud it sits, leave ducking on so it drops under the speech, preview the mix, and export when it sounds right.
What does ducking mean?
It means the music automatically gets quieter while someone is speaking and returns in the gaps, instead of sitting at one fixed level. Here it dips the music 10 dB under speech by default, adjustable from 0 to 30, and it ducks fast but recovers slowly so it does not jump between sentences.
Can I loop a short music clip under a long recording?
Yes. Tick loop and the track repeats to fill the recording, with a short crossfade at each repeat so the joins are not audible. You can also let the music run past the end of the voice if you want it to play out.
Can I mix more than two tracks together?
Not at once. The editor handles one main recording plus one overlay. You can mix down and then add another layer on top of the result, which covers most cases, but a genuine multitrack session needs a desktop editor.
Can I use any music I like?
Technically the tool will mix whatever you load. Legally, publishing something with commercial music in it needs a licence from the rights holder, and that is true no matter which editor you use. Production libraries and Creative Commons tracks exist for exactly this.
Is the music uploaded anywhere?
No, and neither is the recording. Both files are decoded and mixed in your browser, so an unreleased episode and a licensed track both stay on your own machine.
Next door
Lay music under a voice
Drop the file in and it opens ready to work on. No account, no queue, and it never leaves your machine.