Vocal Remover (Karaoke Maker)
Turn the singer down, keep the band in stereo, and hear it before you make the file.
Try before you buy.
- Free preview: the first half of the result (up to 30 seconds) at the quality you chose, with a short MySmartCoPilot chime every 10 seconds.
- Locked until you unlock it: download and share.
- Unlock: Pro pass, ₹179 for 30 days or ₹79 for 1 week, a one-time payment that never renews. Unlock: Pro pass, $5.99 for 30 days or $2.99 for 1 week, a one-time payment that never renews.
Ways to unlock shows how to get the full result.
Printing this result is locked in the free preview.
What to make
Lower keeps a little of the voice as a guide and softens artefacts.
Before and after
20 seconds play in a loop, at the same loudness on both sides; the loudest part is picked first.
Works on stereo mixes with the lead vocal in the centre. Reverb and echo on the voice, and backing vocals spread to the sides, stay in the result.
Result
Locked in the free preview. Opens the ways to unlock this result.
Locked in the free preview. Batch runs unlock with a pass.
Locked in the free preview. Query results unlock with a pass.
About the Vocal Remover (Karaoke Maker)
Make a karaoke or instrumental version of a song. In most stereo mixes the lead vocal sits in the centre — exactly the same in the left and the right channel — while guitars, keys and backing parts are spread to the sides. This page finds what is centred, band by band, and turns it down, keeping the instruments on the sides where they are and, if you like, the bass and the cymbals that are centred too. It can also keep only the centre, for a rough acapella, or use the classic phase cancellation (left minus right).
A 20-second before and after plays the loudest part of the song with and without the vocals at the same loudness, and says how much of it was in the centre, so you can hear whether this song suits the method before you make the file. A video keeps its picture: only the sound is processed.
It is signal processing in your browser: nothing is uploaded and no AI model is used. The free preview on this page is the first half of the result (up to 30 seconds) with a short MySmartCoPilot chime every 10 seconds; a Pro pass, or unlocking this one result, gives the whole file.
How to use it
- Choose or drop a stereo song or music video (MP3, M4A, WAV, FLAC, OGG, MP4, MOV, MKV, WebM…). The page checks that it really is stereo and measures it.
- Choose What to make: Karaoke / instrumental keeps the stereo picture, Classic phase cancellation subtracts one channel from the other, Acapella keeps only the centre. Keep the bass below 120 Hz so the bass and the kick drum stay.
- Press Play under Before and after and switch between Original and Without vocals. If some of the voice is left, try a wider Centre width; if too much music goes with it, a narrower one or a lower Strength.
- Choose the level and the format under Save as, then press Remove the vocals: without a pass you get the free preview to listen to; with a Pro pass, or once this result is unlocked, Download or Share gives the whole file.
Examples
song.mp3 · stereo, lead vocal in the centre · Karaoke / instrumental · strength 100% · keep the bass below 120 Hz
song-karaoke.mp3: the lead vocal much quieter, the band still in stereo with its bass and kick drum, at the loudness of the original. Some of the voice’s reverb is still audible.
music-video.mp4 · Karaoke / instrumental · the same kind of file
music-video-karaoke.mp4 with the picture copied unchanged and the centred vocal turned down in the sound.
track.wav · Acapella · leave out the bass below 120 Hz
track-acapella.wav (mono): the centred voice, with whatever else was mixed in the centre — good enough to try ideas, not a studio stem.
How it works
Mid and side. A stereo song can be written as a mid signal, M = (L + R) / 2 — what both channels share — and a side signal, S = (L − R) / 2 — what differs between them. A vocal mixed exactly in the centre is all mid. The classic phase cancellation keeps only S (left minus right), so the vocal cancels — and so does everything else in the centre, including the bass, unless the mid’s bass is added back, here through a 4th-order Linkwitz–Riley low-pass at the frequency you choose. The result is mono.
Band by band. The karaoke mode works in the short-time spectrum, in frames of about 90 ms. In each band it measures how alike the two channels are over the last 50 ms — their coherence, 2|⟨L·R*⟩| / (⟨|L|²⟩ + ⟨|R|²⟩), which is 1 for a sound in the centre and 0 for a sound on one side only — and turns the centred bands down in both channels. Bands panned to a side keep their level, so the instruments stay where they were in the stereo picture. Centre width sets how far from the exact centre still counts as centred; Strength sets how much is taken out.
Acapella keeps only the centred part of each band — the level both channels share, with the phase of L + R, weighted by how centred the band is — from the bass frequency up to 12 kHz.
What it removes, and what it cannot
- Removed well: a dry lead vocal mixed in the centre, the usual case in pop, rock and film songs.
- Left behind: the vocal’s reverb and echo (they spread across the stereo picture), backing vocals and harmonies panned to the sides, and double-tracked vocals panned left and right.
- Taken along: anything else mixed in the centre at the same frequencies — often the snare, a solo instrument or the bass (keep the bass below 120 Hz to save the bass and the kick drum).
- Not possible: mono recordings and songs whose two channels are the same have no centre to tell apart; the page says so instead of making a file. This page has no AI stem separation.
Tips
- Listen to the before and after on headphones: removing the centre changes the stereo picture, and small speakers hide it.
- For singing practice, a Strength of 70–90% leaves a faint guide vocal.
- Same loudness as the original brings the instrumental back to the song’s level; without it the result is quieter, because the vocal carried part of the energy.
- The classic mode is mono and can sound hollow; the karaoke mode usually sounds fuller.
Limitations
- It needs a stereo mix with the lead vocal in the centre. Mono files are refused, and vocals panned to one side are not removed.
- Reverb and echo on the voice, backing vocals spread wide and double-tracked vocals stay audible; centred instruments at the same frequencies are turned down with the voice.
- The acapella is rough: it keeps whatever was in the centre, not only the voice.
- This page does not use AI separation, which can split mono songs and off-centre vocals into stems.
- Up to 2 hours on a computer and 30 minutes on a phone; files with several sound tracks keep only the main one.
Privacy
On this device; its files download once
Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot. The audio engine (and, when your file needs it, the MP3, AAC or FLAC encoder) downloads from MySmartCoPilot the first time you open a file. Your songs and videos are never uploaded.
Frequently asked questions
What do I get without a pass?
Without a pass, Vocal Remover (Karaoke Maker) shows the first half of the result (up to 30 seconds) at the quality you chose, with a short MySmartCoPilot chime every 10 seconds. Until you unlock it, the result can’t be downloaded or shared. A Pro, Premium or Ultimate pass, a one-time payment that never renews, unlocks the full result. The pricing page lists the passes and their prices.
Is my song uploaded?
No. The song or video is read, processed and written again in your browser, on your own device; nothing about it is sent anywhere.
Why can I still hear the singer a little?
What is left is usually the vocal’s reverb and echo, backing vocals panned to the sides, or a double-tracked vocal: none of them is in the exact centre, so a centre remover leaves them. A wider Centre width takes a little more, at the cost of more of the music.
Does it work on mono songs or old recordings?
No. If both channels hold the same sound there is no centre to tell the voice from, and the page says so instead of making a file. Only AI stem separation can split such recordings, and this page does not offer it.
Will the instruments stay in stereo?
In the Karaoke / instrumental mode, yes: only the centred bands are turned down, in both channels, so the instruments keep their places. The classic phase cancellation and the acapella are mono.
Can I make a karaoke video?
Yes. Open the music video and keep Save as on the same kind of file: the picture is copied unchanged into the new file and only the sound is processed and encoded again (AAC, or Opus in WebM).
What does the free preview include?
Without a pass the page processes the first half of your song, up to 30 seconds, in the format you chose, with a short MySmartCoPilot chime every 10 seconds, and plays it here; for a video, the first half up to 10 seconds with a watermark on the picture. The before and after works too, with the chime on the processed side. The whole file needs a Pro pass, or an unlock of this one result, which makes it from the file still open on the page.
Open-source components
This page can load these open-source components, each only when a task needs it:
- It uses FFmpeg under the LGPLv2.1 and the LAME MP3 encoder under the LGPL, through libav.js: licence and source code.
- It uses FFmpeg under the LGPLv2.1, through the Mediabunny add-ons: licence and source code.
- It uses the LAME MP3 encoder under the LGPL, through the Mediabunny MP3 add-on: licence and source code.
- It uses Mediabunny under the MPL 2.0: licence and source code.
You may replace them with your own builds. Their licences and the credits of every library the site uses are on the open-source credits page.