How to Remove Vocals from a Song in GarageBand

To remove vocals in GarageBand, first check whether you have the original project with separate vocal tracks. If you do, mute the relevant singing tracks and any vocal-only effects returns, then play and export the remaining arrangement. If you only have a finished stereo song, its singing and instruments are already mixed together; use a separate source-separation or center-cancellation workflow, then import the result into GarageBand.

A separator estimates sources from a mix. It can leave singing behind or alter instruments, and its output is not the original session. For the best available backing, compare an authorized instrumental or original project before processing a finished file.

Choose the right starting point

What you have	What to do	What to check
GarageBand project with separate vocal tracks	Mute the lead, doubles, harmonies or vocal-only returns you do not want.	Listen for shared reverb or another unmuted vocal track.
Authorized instrumental	Import it as audio into the project.	Check its intro, key and start position before recording over it.
Finished stereo recording	Test source separation or, for suitable stereo material, center cancellation outside GarageBand.	Check for vocal bleed, missing bass or drums and new artifacts.
Voice rather than backing wanted	Use the vocal output from a source separator, not the remaining mix.	Check consonants, quiet phrases, harmonies and instrument bleed.

Use a recording you own or have permission to process, and obtain the rights needed for any public performance, remix or distribution. Removing a voice does not itself grant those rights.

Remove vocals when you have the original GarageBand project

1. Open a copy of the project and play a section where the vocals and instruments are both audible.
2. Identify the tracks containing the lead vocal, backing voices and vocal effects. Mute the parts you do not want, then listen through a verse and chorus.
3. Check sends and shared effects. Muting a vocal track may not remove every vocal echo or reverb if the routing includes another active path; preserve effects that instruments also need.
4. Export a test and play the actual file. On Mac, Apple documents Share → Export Song to Disk in its GarageBand export example. Check the export range and which tracks are audible.

If the project has only one stereo region containing the entire finished mix, muting that track mutes the instruments as well. The original parts are not exposed simply because the song is on a GarageBand track.

Separate a finished song, then import the backing

1. Choose a supported saved file. Keep the original unchanged and open the Saturalabs vocal remover. For a supported recording, describe the singing you want to remove.
2. Compare both outputs. Background audio is the estimated remainder to audition for a backing; Isolated sound is the estimated singing. Listen to a quiet verse, busy chorus, harmonies and the ending.
3. Download the useful result. AI downloads require an account and sufficient credits, with the cost shown before download. The result is an audio file, not a GarageBand project or a guarantee of a perfectly vocal-free track.
4. Import the downloaded audio. Follow the Mac or iPhone steps below. Keep the original on a separate muted track if you want a timing reference.

For a limited free alternative on a compatible stereo MP3 or WAV up to five minutes and 50 MiB, the center-cancellation test subtracts matching left and right content locally. It can reduce a centered vocal, but can also remove centered bass, drums or other wanted audio; panned voices and reverb may remain. Its mono result is not an AI-isolated instrumental. Compare before using it.

Import the result into GarageBand on Mac

1. Open the project or create an empty project, then drag the downloaded audio file from Finder into the Tracks area. Apple demonstrates this method in its GarageBand for Mac instructions.
2. Place the region at the intended start. If you are replacing a song already on the timeline, line up its beginning with the original region and mute the original so both versions do not play together.
3. Play an obvious transient near the beginning and another near the end. A shifted or shortened file can fall out of alignment even when the opening sounds correct.
4. Save the editable project separately. When ready, export the desired range and listen to the exported file, including its beginning and ending.

Import the result into GarageBand on iPhone

1. Save the authorized audio file to Files. In GarageBand, set the song section length to Automatic if you want to import the whole file.
2. Open Tracks view, tap the Loop Browser, choose Files → Browse items from the Files app, and select the audio.
3. Drag the file into Tracks view and align its left edge with the intended bar or beat. Listen through the end and extend the region or section if necessary.

Apple’s iPhone import guide says that otherwise a region may be trimmed to the current section and that imported audio is converted to 44.1 kHz, 16-bit format when needed. It also notes that imported audio other than Apple Loops does not follow later GarageBand tempo changes automatically. Check timing after changing the project tempo.

Common questions

Can GarageBand remove a singer from a mixed song by muting a track?

Only if that singer is already on a separate track in the project. Muting a finished stereo mix silences everything on that track.

How do I keep the vocal instead of the instrumental?

If you have separate original vocal tracks, export the desired vocal parts. For a finished mix, test a separator and review its vocal or Isolated sound output. Listen for missing words and instrument bleed.

Why did the imported song end early on iPhone?

The current GarageBand song section may be shorter than the imported audio. Apple recommends setting the section to Automatic for a full import; inspect the region and extend the section or region if needed.

Can I use the result for a karaoke video?

With the necessary rights, you can bring a suitable backing into a video editor and add timed lyrics separately. See how to make a karaoke video; the Saturalabs audio tool does not create the video or lyrics.