Review: Audionamix XTRAX STEMS - ProSoundNetwork.com
Audionamix brings its vocal-extraction software to the DJ/pro-sumer market with XTRAX STEMS

Over the years, Paris, France-based Audionamix has made a name for itself with an array of software offerings that can extract voices and instruments from mono and stereo recordings. It’s easy to see why: The idea of reverse engineering an audio track into the sum of its parts is pretty enticing and has a lot of potential uses. Audionamix’s ADX line of software comes in a variety of editions, each designed to cover different applications, but they all have three things in common: They require skill, patience and money to make the most of them. For users without some (or any) of those resources, however, the company has now introduced XTRAX STEMS, an entry-level edition for $99, aimed directly at DJs, remix artists and producers—and it’s a lot of fun.

A standalone application for the Mac only, XTRAX STEMS is fully automated, allowing users to breakdown each track into three stems—Drums, Music and Vocals—even if they have no pro audio skills. After a painless software installation and authentication, users simply drag-and-drop a track on to the software’s dashboard to start the process. XTRAX STEMS works with .wav files up to 96 kHz/32-bit; other formats that it accepts—.aif, .aiff, .aac, .mp3 and .m4a—get automatically converted to .wav.

XTRAX STEMS offers four separation algorithms to choose from: Automatic, which separates vocals only when detected; Generic, which will put the track’s main melodic content and vocals all into the Vocals stem; and Automatic HQ and Generic HQ, which are alternate algorithms for their respective counterparts.

XTRAX STEMS uploads the file to the Audionamix website, where the company’s servers do the heavy lifting; soon, the parsed stems appear and you can dig into the results immediately with the basic tools provided, soloing, muting, panning and controlling the volume of the individual stems. If you’ve run a track multiple times through different algorithms, XTRAX STEMS lets you click through them while the track plays, allowing you to compare results. Users can opt to export an entire track or specific time selection, and the stems are saved as 44.1 kHz/16-bit .wav files.

As it’s entry-level software, XTRAX STEMS provides a down-and-dirty separation in only a few minutes, so there’s often artifacts in the results. More advanced versions of Audionamix’s software offer onboard refinement tools to improve a vocal separation, but those are absent at this price point.

The results can be impressive, depending on the source and what you want to accomplish. XTRAX STEMS separated the drums and vocals—both rapped and sung—on Guru’s classic hip-hop track, “No Time To Play,” with results that were relatively DJ-ready, while on the EDM-oriented “Magic” from Simple Minds’ new Walk Between Worlds album, the vocal was pulled out less cleanly, bringing along much of the melody in between syllables.

Regardless, the software consistently peels away enough material that you get a far greater sense of what’s at work within a mix, making it a useful tool for students of production or musicians learning a song. XTRAX STEMS won’t supplant its more extensive older siblings, but it may well become a stepping stone to them, converting dabbling users into full-fledged fans who want more control over results, leading them to the ADX line accordingly.

Of note: Literally as I was hitting ‘save’ on this review, Audionamix announced V1.1 of XTRAX STEMS, adding compatibility with Native Instruments' Stems file format, stability improvements and workflow enhancements. That version has not been reviewed here, but I never ran into stability issues while using V1.

Audionamix • http://audionamix.com