How to make an acapella from a song

The separation takes one upload. The part that decides whether you can actually play the result is everything after it.

1. Start from the best source you have

Separation quality is capped by what you feed it. A lossless file or a 320 MP3 gives a noticeably cleaner vocal than a 128 rip, because the model cannot recover detail that was thrown away by the encoder. If you own the record in a lossless format, use that one.

2. Separate the vocal

Upload the track. The vocal and the instrumental come back as separate files. The first 30 seconds are free, which exists so you can answer the only question that matters — does this particular vocal survive the split — before spending anything.

3. Judge the result on the chorus, not the intro

Vocals separate cleanly over sparse arrangements and badly over dense ones. An intro will flatter almost any tool. Skip to the busiest part of the record; that is where artefacts live, and that is where you will hear whether it is good enough to play.

4. Get the key and tempo right — on the full mix

This is the step most guides skip, and it is the one that decides whether the acapella is mixable. Do not analyse the separated vocal: with no kick and soft transients, beat trackers perform far worse on a bare vocal than on a full record. Take the key, the tempo and the downbeat from the original mix and apply them to the stem.

Here that happens automatically — both halves inherit the analysis of the full mix.

5. Import it gridded

Serato stores its beatgrid inside the file's tags, so a properly tagged AIFF opens ready to play. Files that arrive untagged have to be analysed and often hand-gridded, per file, forever. See acapellas for Serato for the detail.

Questions

Related