Everything You Need To Know About The New Stem Audio File Format

There's a new kid on the block in the audio world, targeted at music performers, called STEMs. Professional audio folks should already know all about it, but the quiet hobbyist might still be unaware of just how much the mighty Stem will change their life.

Every STEM file features five channels. There's a stereo channel to play the track as normal, through any basic speaker system. Then there are four separate channels, usually dedicated to drums, basslines, melodies, and vocals. How one uses the different channels is up to them, but the aforementioned format is conventional in the music space — handy for when you have different knobs or sliders mapped to different channels on your hardware.

There will undoubtedly be many creative uses of the format. We'll get into that below. But STEMs will have the greatest effect on DJs, producers, and mashup artists. For those groups, this is a dream come true.

Previously, DJ mixers allowed you to manipulate the low, mid, and high portions on the track EQ. That meant elements were getting bundled into one another — attempt to bring down the vocals, and you'd be bringing down the high hats as well. Try to get rid of the bass, and you'd inevitably be cutting the kickdrum too. We did the best we could with it, but it was basically faking the level of control that STEMs offer.

Now, you can cut back and forth between vocals on a house track easily. Stick one rapper's lyrics over a different hip hop beat with ease. Switch the basslines of two tracks while the other elements play normally, for a seamless and impressive-sounding mix. It's a game changer.

What about Serato?

Seeing as the STEM format was created by the folks behind Traktor, that performance software has obviously enjoyed a bit of a head start. Still, Serato has been slow on the uptake and doesn't yet support STEMs. Traktor was already ahead according to most DJs I speak to, and this is only widening the gap.

Do I need special hardware?

Nope!

There are already bits of DJ hardware coming out purely for STEM manipulation, and it's not too hard to repurpose other controllers to work with STEMs. Traktor products like the Kontrol F1 for $249 are plug & play, or the $699 Kontrol D2 is pretty much made with STEMs in mind. There's also the $1449 Kontrol S8 as an all-in-one solution.

Those last couple of options are a little pricey, though. Something like the Midi Fighter Twister would do everything you want to do — here's a video of that in action. Anything with mappable buttons, knobs, or sliders would do, such as this $121 Behringer, or this $165 version.

But you don't actually need any of that. With Traktor being customisable and all, you could just map the STEM controls to your keyboard. It'll take some thinking — you'll need to find eight keys you're happy to dedicate to STEMs, that will also be clear enough for you to not hit the wrong button. Perhaps some sort of coloured labelling is in order. With that solution, you're basically only getting cut-offs, whereas the dedicated hardware gives you knobs to gradually increase or decrease the volume of a channel.

So obviously if you intend to use this professionally, you'll want to have the appropriate hardware. But for the hobbyist, even the limited control of cut-offs is still a lot more control than you had with just lows, mids, and highs. You'll be able to cut the bass on one track and bring it in on another, adding a fun tool to your DJ arsenal.

Can I make STEMs?

Yep!

It's an open file format, so there are no licences needed or royalties that need to be paid. You can't rip audio channels out of existing MP3s to make a STEM file, but you can absolutely combine audio files to make one. I imagine some folks will get pretty creative with the possibilities there.

There's a freely downloadable Stem creator which allows you to pop everything in, make sure everything is lined up properly and the levels are right, edit the metadata, and press go. Bring in your project from Ableton Live, or do whatever other crazy idea you've got in your head.

Just as one example, if you've got the a capella as well as the instrumental of a track, you could absolutely combine the two to make your own mini-STEM — regardless of whether the producer has cottoned onto STEMs yet or not.

The creator has a great tutorial video on how to use the app:

Is it lossless?

It can be. According to the STEM FAQ:

The STEM format uses the mp4 framework which allows STEM files to be encoded either in AAC 256 kbps VBR or Apple Lossless audio (ALAC). The upcoming Windows 10 provides native ALAC support however previous Windows operating systems do not. Please note that any playback software used on older Windows systems needs to support ALAC itself.

Where can I buy them?

Online marketplaces dedicated to DJing - such as Beatport - have been quick to see the potential. Find them there, as well as Bleep, Juno, Traxsource, whatpeopleplay, Wasabeat, and whoever else sees the light in the near future.

The availability of STEMs relative to your favourite tracks depends entirely on the producer, though. Most electronic genres will benefit greatly from it, so it's easy to see producers seeing this as an easy way to increase value offered and reward for effort. Lower level producers might take longer to start releasing them, but from Hip Hop to House to Dubstep, this rising tide will lift all boats.

Additional uses

There's no telling what people will come up with, and I'm betting we'll see the next Madeon in a little while on Youtube, amazing us with ideas we wish we thought of, with new hardware dedicated to the format. That's in addition to the more routine stuff like "Boy I wish I had that track without the vocals."

Maybe podcasters or performance artists will do something interesting with it. All that's required is the listener having an adequate way to raise or lower different volume channels.

Have you ever listened to a music podcast and wanted the host to shut up so you could listen to the music? I know electronic music fans can attest to the desire to strangle anyone who talks over the drop. Like radio, a lot of them talk over the music so you can't rip the tune, but with STEMs it'd be possible to just turn the unwanted element down. Bye bye, chatty DJ. We won't miss your shouts out to people we've never heard of.

Or perhaps it's a talking podcast, something to do with news or history, and you want to hear a little more about what was just mentioned. Flip over to Channel 2 where they go more in-depth, Channel 3 where they have an extended interview with the relevant expert, or Channel 4 for the bloopers. Maybe it's a four person chat podcast and you hate the fourth person. Bye bye, fourth person.

Maybe you'd like to switch between different commentary or languages while you listen to a sports show?

One could imagine for ages. But one of the coolest aspects is that you can try it out for free. Well, almost free. After you've bought a few STEM tracks, you can go ahead and map some keys and try it out on your existing software. If you get inspired, then there's the dedicated hardware to think about. Perhaps it's not too late to ask friends and family for a combined Christmas gift...


Comments

    One idea for potential: Rhythm games. Like how in Guitar Hero if you mess up the guitar track goes silent. Seperate STEM tracks for Guitar, Bass, Drums, Vocals (and one for everything else), each one going silent or distorted if you mess up.
    Of course, I'm sure you could get more creative than a simple Guitar Hero / Rock Band clone..

      I'm sure they have that already on some of the games?

      If you mess up you get the distorted crunch, but if you miss notes on say, the guitar track, the bass drums and vocals continue as normal?

      Or you could just play rocksmith where it plays whatever you play.

    This doesn't seem to be a new format, but just a standard for assigning different music parts to different channels of an mp4 container.
    A FLAC file can have up to 8 channels so there shouldn't be any reason it couldn't be adapted to work with that too.

      Yep. Multi-channel formats have been around for ages

      To my knowledge, you can't do the same things mentioned above with FLAC, though I'm not sure of the specifics of why. DJs who use FLAC aren't manipulating channels on the fly, they're just going for lossless.

        I wouldn't be surprised if, soon after stem-aware hardware becomes more readily available, people smarter than me will create software allowing that hardware to work with other multi-channel containers.
        MKV would be nice, because in theory it would allow for any combination of codecs to be used together.

        Still seems strange to me that whoever came up with stem decided to limit it to AAC.

          Honestly if it were going to happen, I think it would have by now. Basically any MIDI controller on the market worked with Stem right off the bat -- them, as well as Traktor and Serato are highly customisable and mod-friendly, and all of the above take FLAC so it's not a prioprietary issue. So if it were possible to map keys to specific FLAC channels, it would have been done, as it's pretty much a DJs dream come true.

          I haven't been able to find anything along those lines while searching, and I've never known a DJ to do it. Though I think you're right, it's odd that it hasn't happened yet. I'd be interested to hear the technical reasons why starting a new open format was preferable to using FLAC, and why devs/modders hadn't solved that problem before now.

          I'm not complaining though, just really happy it's here :)

    Pretty sure FLAC is far superior to this and has been for a long time....

Join the discussion!

Trending Stories Right Now