It’s a humbling phenomenon: just when you think you're getting a handle on the daunting world of mixing, a new technique comes along and knocks you back to square one. For me, stem mixing was one such technique. Tracks are what we’re mixing, right? What is a stem and why should I care?
In simple terms, stems are mini-mixes: manageable groups of like instruments or other elements that can be manipulated in broad strokes, allowing changes to be made without going back to the drawing board every time you need to tweak a mix. After all, people often ask for simple changes — brighter vocals here, less piano there. Rarely do people say, "I'm hearing a little too much compression at 6k on the 3rd harmony guitar."
Most small projects probably won’t benefit substantially from this technique, but it’s used all the time in the upper echelons of the industry and not without reason. Stems, for instance, facilitate collaboration by keeping things simple when a project changes hands.
Stems facilitate collaboration by keeping things simple when a project changes hands."
They’re also a standard part of TV and film production, where mixers often deliver the separated elements of dialogue, music, and sound effects (DME in industry jargon). Even games like Rock Band and Guitar Hero use stems to keep the rhythm section going while you flub the guitar solo or take out the vocals so you can embarrass yourself karaoke-style.
So why should this matter to you? Why add an extra step and convolute an already complex process like mixing? Plenty of reasons:
The artist wants to be involved in mixing, but you don't want to overwhelm them with a complex session (or let them screw anything up).
You're an electronic producer whose session is approaching the 100-track mark, and you need a slimmed-down version to mix.
You went a little overboard on resource-hungry reverbs and effects, and your computer's processor can't keep up.
You trust your mastering engineer and want to give them a little more control over your track without having them mix every element for you.
The song you're mixing is being licensed for TV or film, and you need to be ready to present instrumental or a capella versions on demand.
If you’re at all serious about mixing professionally or just want to expand your capabilities, it's a worthy technique to know about. So pull up an old mix session and follow along as we explore the basics of stem mixing.
Making Your Own Stems
First, just mix the song like you normally would. Get everything sitting right, check it on a few different speakers and headphones, then go back and tweak a few times.
Pay more attention to tone and balance than the overall levels at this stage. Make sure the drums are playing nicely together, the rhythm and lead guitars compliment each other, vocals blend smoothly, etc. Save the DAW session so you've got a clean slate to go back to if anything goes astray from this point on.
A word of warning here: Don't put anything except a mild peak limiter on the master bus. A master compressor might make your mix sound fat, but it will react very differently when you mute, solo, or re-mix your stems. If you normally mix with a compressor on the output, save yourself a headache and take it off and compress individual elements instead.
Once your mix is solid, save the session again under a new name (something like "Songname-stems" would make sense). Now, start by making a new stereo track for each of the stems you want to create. There can be as few or as many as you think you need. It's a balancing act between simplicity and control. Always name the tracks appropriately, as this will determine the names of the actual audio files you create later.
A good starting point for a rock mix might be drums, guitars, bass, vocals, and vocal effects stems; an electronic mix can be broken up by percussion, rhythmic synths, melodic synths, pads, and effects; orchestral stems could include room mics, conductor mics, soloists, and each of the various instrument sections.
Assign the inputs of your new tracks to internal stereo busses (bus 1/2, 3/4, etc.), then route the outputs of your original tracks to those same busses (instead of the main outputs), depending on which stem group they belong to. Don't change any levels, effects, or panning. The idea is to print everything as-is. As long as the new tracks’ faders stay at 0db, your mix should sound exactly the same.
You could stop here, minimize the original tracks, and have a nicely simplified mix but retain the ability to make minute tweaks, but true stem mixing involves one more step: recording the actual audio for the stems. There are two ways to do this.
Don't change any levels, effects, or panning. The idea is to print everything as-is. As long as the new tracks’ faders stay at 0db, your mix should sound exactly the same."
If you want to be able to mix the stems yourself, simply arm the stem tracks, hit record, and let the song play through. Let it roll for a few seconds after the last sound dies away, hit stop, and you've got stereo tracks for each of your stems.
Now, you're free to hide, disable, or even delete the originals. Hiding and disabling is a smart move so you can free up processing power but still be able to re-activate them in case you need to make changes and re-print a stem.
If you don’t need the stems yourself but are instead sending them off for mixing, mastering, or licensing, you could export them (or bounce or render, whatever your DAW calls it) instead of recording them internally. It’ll take a bit longer, but you’ll get the exact same files, and you can easily save them to any location instead of having them clutter up your project.
Go to the same export menu you’d use when bouncing the final mix of your song, but instead of choosing the master output, select one of your stereo stem busses. Name the file, export it, and repeat the process for each stem.
Now, you have a slimmed-down set of elements to manage, while still having control over overall levels and tones. Drums need a boost to keep up the energy? Synths a little too bright? Want to try a version with dry vocals? All these tweaks are easy to accomplish with stems.
Further Exploration
Beyond the conveniences of stem mixing, there are a few other applications for the technique.
Stems make it easy to add analog sound to your digital mixes. You can do all of the editing and detail work in your DAW, then send your stems through an analog console to impart a cohesive sound without setting up a huge, Pink Floyd-style board mix.
Conversely, you can print them to tape to add some organic saturation instead of doing the entire project that way. Some engineers even run tracks through tape machines without recording just for the sound of the circuits.
They also open a world of possibilities for re-mixing. Having access to separated elements allows a producer to take out and replace instruments, pull clean, isolated samples, or rearrange the song with ease.
Start making stems a part of your workflow, and you’ll be efficient, flexible, and ready for any situation.