Vocal Remover

Vocal Remover makes a karaoke version of a song by cancelling the sound that sits in the middle of the stereo mix. Most pop, rock and dance tracks put the lead voice dead centre, equally loud in the left and right speaker, while guitars, keys and backing parts are spread to the sides. Remove what both sides share and much of the voice goes with it.

Plain “left minus right” cancelling also wipes out the bass and kick drum, which are centred too. This tool splits the song into frequencies first, so it can leave the low end and the very top untouched, and its centre-focus control only removes sound that is truly the same in both channels. That keeps more of the music than the classic trick. A “vocals only” mode does the opposite and keeps the centre, for a rough acapella.

Be clear about the limits: this is a quick, maths-only method, not an AI separator. It needs a stereo song with a centred voice. Mono recordings, live tapes and songs with doubled or wide vocals can’t be cleaned this way, and some reverb usually remains. An AI mode may come later. Everything runs inside your browser, so your music is never uploaded.

1. Tap “Choose a song” and pick an MP3, M4A, WAV, FLAC or a video with music.
2. Press “Play” and switch between “Without vocals” and “Original” while it plays to compare. Tap the waveform to jump to a chorus.
3. Adjust Strength, Centre focus and the two “Keep” sliders, or try the Quick settings, until the voice is as faint as you can get it.
4. Choose “Save as” WAV or MP3 (other formats work too where your browser supports them).
5. Press “Save karaoke track” (or “Save vocals” in vocals-only mode), then “Download”.

• Centre cancellation done per frequency, not on the whole sound at once.
• Keeps bass and kick drum: low notes below your chosen point pass straight through.
• Keeps cymbals and air: very high frequencies can be left alone too.
• Centre focus from classic (cancel everything centred) to strict (only identical sound).
• Vocals-only mode for a rough acapella, clearly marked as rough.
• A/B listening: flip between the result and the original at the same moment.
• Warns you straight away when a file is mono and can’t be separated.
• Saves WAV or MP3, with an option to bring the level back up after cancelling.

• If the voice still comes through, lower Centre focus a step at a time: it removes more, at the cost of some centred instruments.
• If the drums sound thin, move “Keep bass below” up to 150–200 Hz so more of the kick stays.
• Reverb on the voice is usually spread to the sides, so a faint echo of the singer often remains. That’s normal for this method.
• Use the best-quality copy you have. A low-bitrate MP3 or a video soundtrack often smears the stereo picture and leaves more voice behind.
• Old songs with the singer on one side, and most live recordings, don’t have a centred voice, so they won’t work well here.