/
Home

Categories

Network ToolsText ToolsGeneratorsImage ToolsPDF ToolsAudio ToolsDevice TestsCalculatorsTime ToolsDeveloper ToolsColor Tools
About

Vocal Remover

Split a stereo song into an instrumental and a vocal/center track.

  1. 1
  2. 2
  3. 3

Key features

Instrumental (vocals reduced) or center/vocal track
Preview both versions before downloading
Upfront about what this technique can and can’t do
100% private — processed locally, never uploaded

Upload a stereo song to get two versions: an instrumental track with center-panned vocals reduced, and a center/vocal track isolating what’s panned dead-center — most songs mix lead vocals there, so this often approximates a karaoke track and a vocal-only track.

This uses phase cancellation (subtracting the right channel from the left, and averaging them), a decades-old audio trick — not AI source separation. It works best on studio recordings with vocals mixed dead-center, and won’t do much on mono files, live recordings, or songs where vocals share a stereo position with other instruments. Everything runs locally using the Web Audio API; your file is never uploaded anywhere.

How to use

  1. Upload a stereo song

    This technique needs a stereo file — mono recordings can’t be split this way.

  2. Choose instrumental or vocal

    Preview each and pick whichever sounds best for your file.

  3. Download the result

    Export the selected version as WAV or MP3.

Frequently asked questions

How does this actually work?

Most studio recordings pan lead vocals dead-center, with instruments spread across the stereo field. Subtracting the right channel from the left cancels out anything identical in both — usually the vocals — leaving mostly the instrumental. Averaging the two channels does the opposite: it isolates what’s common to both, which is usually the vocals (plus any other center-panned elements like bass or kick drum).

Why didn’t it remove the vocals well on my file?

This only works when vocals are panned dead-center and separated from other center-panned sound. It won’t work on mono audio, live recordings, or songs mixed with the vocal slightly off-center — and it can’t fully isolate vocals from a bassline or kick drum that share the same center position. This is a real limitation of the technique, not a bug — true AI-based source separation does better but needs a large model, which isn’t something this lightweight, fully client-side tool uses.

Is my song uploaded anywhere?

No. Everything is processed locally in your browser using the Web Audio API — your file never leaves your device.

Link copied