Vocal Splitter
With the Vocal Splitter you can extract the vocals and instrumentals of any of your tracks. This means that you can create new tracks without the voice or with only the voice of a track.
The Vocal Splitter uses advanced machine learning to accurately extract vocals from any track. This is why it works much better than tools that simply isolate frequencies.
If you need acapella's for your mashups or maybe you need a karaoke version of a track, the Vocal Splitter is perfect for the job.
The Vocal Splitter uses Spleeter under the hood. Spleeter is currently the best and most accurate way to extract vocals.
Sampler
You can also create samples from within the Vocal Splitter. Just open the Sampler and drag anywhere to create a sample.
Samples are automatically added to a special rekordcloud playlist.
Output
Everything you save is also saved to a special rekordcloud playlist. The next time you download your library, you'll see your new vocals and instrumentals there.
MP3's are encoded as 320kbps.
You can save as WAV files but it's not recommended because WAV does not support all tags and takes up much more space.
Hotkeys
The following hotkeys are available when you're using the Vocal Splitter:
| Key | Function |
|---|---|
Backspace | Goes to the next track without saving |
1 | Sets volume of the original track to full and mutes the others |
2 | Sets volume of the vocals to full and mutes the others |
3 | Sets volume of the instrumental to full and mutes the others |
F | Flips the volume between the original track and the vocals+instrumental |
S | Stops playback |
Enter | Save |
Space | Play/pause |
Scroll | Scroll on a volume bar to increase/decrease volume |
Shift+Click | Plays a track and turns on volume for only that track |
Click Sampler | Set start point of sample |
Shift+Click Sampler | Set end point of sample |
Limitations
The M4A and AIFF file types are supported but have no original track preview.
The vocals and instrumental tracks (and samples taken from those tracks) are limited to 16kHz. This is because of performance and memory reasons. Samples taken from the original are not limited.
Output may contain "artifacts". These strange sounds are a side effect of the vocal extraction and can't really be prevented.