Neural Mix Pro moved DJ-style separation into its own app

The tool moment

On 27 August 2020, Algoriddim launched Neural Mix Pro, described in its own press release as "the first AI-powered music player and editor for producers, DJs, and musicians." The release states the app "fully leverages Apple's Core ML technology to separate vocals, drums, and instruments in real-time," letting a user isolate or mute any of the three directly from a music player rather than inside DJ software. This followed, by just over two months, Algoriddim's original Neural Mix announcement, which introduced the underlying separation technology inside djay Pro AI on 18 June 2020 and called it "the world's first real-time vocal and instrumental separation."

What the documents show

Both press releases are Algoriddim's own claims about its technology, including the "world's first" and "patent pending" language in the original announcement, which this note treats as a vendor position rather than an independently verified priority claim. A later press release, dated 6 May 2021, documents that Neural Mix was subsequently expanded to "four channels of music isolation," separating "the vocal, harmonic, bass, and drum components of a track individually, in real-time," a stem-count increase from the three named in Neural Mix Pro's original announcement, and evidence that Algoriddim treats the underlying model as something it continues to revise rather than a fixed 2020 release.

What stays with the musician

Because Neural Mix Pro is framed as a standalone editor rather than a DJ-set tool, a producer or musician using it still decides what an isolated stem is for: sampling, practice, or building a new arrangement from an extracted vocal or drum part. Algoriddim's press materials describe the separation mechanism and its expanding stem count but do not address whether a given use of an isolated part, releasing a remix built from it, for instance, clears any rights the original recording carries, a question that sits outside what either press release documents.

Judge it by listening

Algoriddim's own releases assert sound quality in general terms, "pristine sound quality" and "sample-accurate precision," without publishing a listening-test methodology. This is an editorial suggestion: since the company's own documents record a jump from three channels to four between August 2020 and May 2021, a useful comparison is the same track's isolated drum or bass channel before and after that expansion, if older app versions remain available, to hear whether adding a channel changed the separation quality of the others.

• Has the four-channel Neural Mix actually improved separation on the channels that existed before the 2021 update?
• Does an isolated vocal or drum part from Neural Mix Pro hold up outside the live-mixing context Algoriddim demonstrates?
• What rights question would using an isolated part in a released remix or sample actually raise?

Algoriddim's own press archive gives an unusually precise, dated record of how Neural Mix moved from a DJ-software feature into a standalone editor and then grew a fourth channel, even though the "world's first" framing in that record is the company's claim to evaluate, not an independent finding.