Turn Any Song into Karaoke

How it works

A singer’s karaoke workflow

I’ve spent 20 years mixing and mastering records in Melbourne and a decade teaching audio, and the singers who sound best on the night are the ones who rehearsed with the real arrangement. A karaoke track is a rehearsal tool first and a performance track second. This is the workflow I give students.

1. Start with the best source file you have

The separation can only work with what is in the file. A lossless download (WAV or FLAC) or a high bitrate MP3 or M4A of the original master gives the cleanest instrumental. A low bitrate file ripped from a video has already thrown away detail the AI uses to tell the voice from the band, and you will hear it as a slightly watery top end. If you own the song on CD or as a purchased download, use that.

2. Split with Vocal Remover

Choose Vocal Remover for a karaoke track. It gives you two stems: vocals and instrumental. That is all a sing-along needs. If you also want to drop the guitar or piano to play the part yourself, Quick Split (4 stems, free) or Full Separation (6 stems, Pro) break the band down further.

3. Keep a guide vocal while you learn

This is the step most karaoke tools skip. The split opens in a live mixer, so instead of muting the vocal, pull its fader down until it sits just under the band. You still hear where the phrases start and how the melody moves, but your own voice is on top. As the song settles in, bring the guide down further each run until it is gone. Then bounce the instrumental on its own for the performance.

4. Check the chorus for ghost vocals

Before you commit, solo the instrumental and skip to the biggest chorus. That is where doubled vocals, harmonies and long reverb tails live, and it is where any faint vocal residue will show up. In a verse you will rarely hear anything. If the chorus is clean, the whole track is clean. If you hear a trace, try a better quality copy of the song.

5. Put the song in your key (Pro)

Most people who say they cannot sing a song really mean it is pitched for someone else’s voice. With Pro you can change the key without changing the speed, a semitone or two down for a lower voice or up for a higher one, and the groove stays exactly where the band played it. Rehearse in the new key so your ear locks onto it.

6. Slow it down, count in, loop the hard bar

Fast verses and wordy bridges are learned slowly. Pro lets you slow the song down without changing the pitch, add a count-in so you are ready for your entry, and save named loops. Free includes basic looping, which is often enough: loop the bar you keep fluffing, sing it ten times, then bring the tempo back up.

7. Add lyrics or video elsewhere

StemConsole produces the audio. Lyrics on screen and karaoke video are made in a separate video editor or karaoke app: import your downloaded instrumental there. For the full step-by-step guide, read how to make a karaoke track.

For hosts, parties, church and school performances

Karaoke hosts and parties: build a set of songs no karaoke catalogue carries, from the original recordings your crowd actually knows. Saved sessions in My songs keep every split ready to reopen, and share links let a guest rehearse the track before the night.

Church and worship teams: turn a worship recording into a backing track for a Sunday when the band is short, and move it into the congregation’s range with Pro key change.

Schools, choirs and theatre: give students the real arrangement without a vocal model to copy, keep a faint guide for early rehearsals, and play WAV bounces (Pro) through the hall PA on performance night. Music teachers can loop the section a student is stuck on during the lesson itself.

• Guide-vocal fader, solo and mute on every stem, free
• MP3 downloads and MP3 mix bounces free, WAV with Pro
• No watermarks on any download
• Works in the browser on phone and desktop, nothing to install
• Your audio is not used to train models

Need a bed for a cover video rather than a sing-along? Use the instrumental maker.

Karaoke methods compared

There are four common ways to get a karaoke track. Here is how they differ from an engineer’s point of view.

Phase cancellation is the old trick: flip the polarity of one side of a stereo file and anything panned dead centre cancels. Lead vocals are usually centre, but so are the kick, snare and bass, so the track comes out thin, and reverb on the vocal stays behind. AI separation learns what a voice sounds like and removes it wherever it sits in the stereo image, which is why the band stays whole.

Frequently asked questions

More tools and guides