1. Upload the song
Choose a WAV, MP3, AIFF, or FLAC file. A clean, high-quality source usually gives the separator more information to work with than a heavily compressed or clipped file.
2. Let AI separate the vocal
VocalRemove analyzes the full mix and creates a vocals-only result. Dense arrangements, long reverb tails, backing vocals, and effects can make separation harder, so treat the preview as the quality check for your particular track.
3. Preview the isolated vocal
Listen in the browser before you commit the result to a remix, mashup, sample, DJ edit, or production session. Previewing is useful because every recording separates differently.
4. Download and use the result
The current pipeline produces a validated 44.1 kHz stereo 24-bit WAV master and a separate MP3 preview. You remain responsible for the rights and permissions required for the song and your intended use.
AI separation versus old vocal-removal tricks
Phase cancellation can work when you have matching stereo mixes with predictable center-panned vocals, but it is sensitive to the mix and often removes other centered sounds too. AI source separation uses learned patterns from the audio to estimate the vocal layer and the accompaniment independently. It is not perfect, but it is a more flexible starting point for finished songs.
What makes a good source?
- Use the highest-quality file you have.
- Avoid files with clipping, heavy re-encoding, or long silence.
- Expect more bleed when vocals overlap with loud instruments or wide effects.
- Preview the result before building a final release around it.