Algoriddim has introduced Neural Mix Pro, an AI-powered music player & editor that’s designed to ‘unmix’ tracks, letting you isolate beats, instruments, and vocals from any song.
The software lets you create beat-precise loops, change tempo, and transpose the key of a song, while removing vocals, percussive, or harmonic parts in real-time. Export allows you to extract a cappellas and instrumentals.
Features:
- Music Source Separation
- Extract beats, instruments, and vocals of any song
- View multi-track waveforms side by side
- AI Music Player
- Isolate beats, instruments, and vocals in real-time
- Mute vocal or instrumental components individually
- High-quality time-stretching: change the tempo of a song without affecting its pitch (up to +-75%)
- Pitch-shifting: transpose a song into a different key (up to +-1 octave)
- Automatic BPM and key detection
- Frequency colored waveforms
- Beat-quantized looping
- Offline Export
- Output options: Instrumental, Acappella, Drums, Harmonic, Harmonic+Vocals, Drums+Vocals, Channel Mix
- Optional target BPM and key transformation for exported files
- Multi-track export
- WAV and AAC output file formats
- Loop Editor
- Extract beat-matched vocal, percussive or harmonic loops individually
- Trim precisely to an audio region or sample
- Set auto-quantized or manual loop in & out points
- Export loops into different file formats using adjustable target BPM and key
- Music Browser
- Music and Finder integration
- Multi-track playlist & folder export of instrumentals, acappellas
- Powerful track sorting and search options
- Built for macOS
- Core ML for advanced machine learning
- Multi-core analysis and export
- Support for Touch Bar
- Multi-touch trackpad control
- Accessibility through VoiceOver
Pricing and Availability
Neural Mix Pro is available for $49.99 USD. It’s free to download and test.
worst demo ever, please tell me you have another demo with better music and a better demonstration to go Algoriddim you should know better!!!
I was going to say “I bet this is based on spleeter” but those isolated vocals sounded worse than anything I’ve heard come out of it.
If this works as promised, I could see it being useful for:
1. stealing others’ work
2. isolating parts for learning
3. various problem solving
4. pitch & time for various uses
The examples we heard are pretty poor quality. Makes me think #2 in my list may be its main use.
Probabily the algorithm used is Spleeter: sometime it work sometime it doesn’t, but it is always interesting to listen to the results. A free online service that I use for this purpose is http://www.ezstems.com
I just finished trying it out with a few songs. With Lucky Man (by ELP), Lake’s voice is isolated quite well: the only issue is, as in the demo, that slight warping in some points, typically where there are more instruments and/or choirs playing. The drum trucks too are precisely isolated, but the sound is terrible: same phenomenon I pointed out for the vocals, but vastly worse. It’s the same kind of sound you get when you transpose a song, if you know what I mean: and, not by chance, drums are exactly the instrument that suffers more from that process. Finally, everything else gets categorised as “Harmonic”. The guitars sound just slightly better than drums, everything sounds like it was deep-fried in flanger and chorus, while Emerson’s Moog solo is well defined per se, but you can still hear some guitars, choirs and drums (strangely enough) in background.
Other songs gave mixed results; some were definitely better (electronic music above all), some even worse (in Burn, by Deep Purple, the instrumental section is incredibly muddy, except for the highest-pitched notes, while Coverdale’s voice is much worse than Lake’s).
All in all, it’s an interesting experiment, but results vary way too much. As using it in a professional setting is unthinkable, one should judge it as a learning/rehearsal tool. Even then, sometimes it’s acceptable, sometimes plain terrible. As there are free services like Songsterr, with huge databases of well-crafted songs divided by tracks that you can isolate, download, print, play in Finale or Sibelius etc, I really couldn’t justify paying for this programme right now.
I’m sure it will get better in the future, there’s however a limit to this kind of apps: the programmers really did their best, but I don’t believe there’ll ever be an AI than can, for example, correctly discern Hammond, guitar and that other layered sound I can’t quite put my finger on it. And above all, the “post-processing”, transposed-like effect (which I suppose is an inevitable drawback), is what ultimately ruins an otherwise impressive, despite its limitations, piece of software
Thanks for that report, Andy. That’s kind of what I suspected.
Requires Mac OS Mojave or later.
Yuck, sounds awful.
if this is the best they could come up with for a demo I dont want to here the rest.