How-to guide
How to isolate vocals for transcription
Transcribing lyrics by ear is hard enough; trying to do it with a full band competing for space in your headphones is brutal. VocalSplit strips everything except the voice so you can hear every consonant.
Isolated vocals are also useful for melody and pitch transcription work — music education tools like Transcribe! and Amazing Slow Downer work much better with a clean vocal stem than a full mix.
Step-by-step
- Upload the song to VocalSplit. Any of MP3, WAV, FLAC, M4A, OGG, Opus or AIFF.
- Pay for the split. $0.99 for one, or a pack if you are transcribing a whole set.
- Download the vocals stem. You only need the vocal stem for transcription — discard the instrumental.
- Load it into your transcription tool. Transcribe!, Amazing Slow Downer, or even the built-in slow-down in any DAW.
- Work through it at slower speeds. Slow the vocal stem to 50 or 75 percent speed — every syllable and pitch becomes clearly audible.
Tips for better results
- For lyric transcription, an AI transcription service like Whisper gives even better results on an isolated vocal stem than on a full mix.
- Melody transcription is much easier when there is no harmonic competition from the instruments.
- Layered backing vocals will come through in the same stem as the lead — separate with a pencil and patience.
Try VocalSplit free
Upload a song and get clean vocals and instrumental stems in under 15 seconds. First split is $0.99.
Split a song