How to Prepare Stems for Mixing

Run through this before you export. Every line is explained further down the page, followed by the export steps for Logic Pro, Pro Tools, Ableton Live and FL Studio. Engineers have their own preferences, so if yours has sent a list, theirs wins.

People use “stems” for two different things. Strictly, stems are grouped submixes: all the drums on one stereo file, all the backing vocals on another. Multitracks are every individual track in the session: kick, snare, overheads, bass, each vocal. Most mix engineers want multitracks, because a mix is built by balancing and treating each part on its own.

Grouped stems make sense when a session is very large, when a producer wants to lock in a sound, or when the music is headed for picture or a live show. If that is what you are sending, say so up front and agree on the groups first.

Either way, send only what plays in the song. Leave out muted tracks, unused takes and ideas that are not part of the arrangement, or put them in a folder clearly marked as extras.

Export uncompressed audio. WAV is the safest choice across every DAW, and AIFF works too. MP3 and other compressed formats throw away detail a mix needs, so keep them for rough mixes you listen to in the car.

Keep the session’s own sample rate. If you recorded at 48 kHz, export at 48 kHz: converting on the way out adds a processing step and gains nothing. Tell the engineer which sample rate you used.

Use 24-bit, or 32-bit float if your DAW offers it. A 32-bit float file keeps peaks above 0 dBFS instead of clipping them, which leaves room to work if a track ran hot. Leave normalizing off, and leave dither for the final master.

Export every track from the same start point, usually bar 1 or the very start of the session, even if the part does not come in until the second chorus. Every file should also end at the same point, a little past the last note so reverb and delay tails can finish.

In most DAWs that means exporting all the tracks together over one range, or consolidating (bouncing in place) each track from the start. Files that begin wherever their first note happens to be have to be lined up by ear, which is a slow and error-prone way to start a mix.

Write down the tempo and key, and any tempo or key changes with the bar they happen at. If the song was recorded without a click, say so, and include a tempo map or MIDI file if you have one.

Some processing is part of the sound of a part: a guitar amp simulator, a synth patch, a crushed drum loop, a vocal effect the song was written around. Print those, because the engineer cannot recreate them without your plug-ins and settings. Print virtual instruments to audio too, and include the MIDI if you can. If you are unsure about an effect, send the part twice, printed and dry, and name them so it is obvious which is which.

Other processing is part of the mix, and belongs to the engineer: anything on the master bus, mix EQ and bus compression, and the shared reverbs and delays that glue parts together. Bypass those before you export. If a reverb or delay is part of the sound, print its return as its own file rather than baking it into the dry track.

Pitch correction sits in between. If the tuning is part of the style, print it. If it is there to fix a take, send the tuned version and include the untuned one as well. Finally, check levels: nothing should clip. If a file does, lower its fader or clip gain and export it again, rather than putting a limiter on it.

“Audio 1_07.wav” means nothing outside your session. Give every file a number and a plain name that says what it is and, where it matters, how it was recorded: 03_Snare_Top, 04_Snare_Bottom, 08_Gtr_Rhythm_L. Keep the numbers in the order the tracks sit in your session. Drums first and vocals last is a common layout.

Put everything in one folder named for the song, with the tempo and sample rate in the name if you like, and keep extras such as alternate takes or printed effects in a subfolder.

For vocals, send the comped lead: one track built from the best takes. Include raw or alternate takes only if the engineer might need them, and name them clearly. If backing vocals are doubled or stacked, keep each part as its own file.

A rough mix is the most useful file you can send. It shows the balance you have been living with, the parts that matter and the moments that should jump out. It does not have to sound good; it has to be honest about how you hear the song.

Add one to three reference tracks: released songs whose sound you like, with a line on what to listen for in each, such as the vocal level, the low end or how wide it feels. References work best as a direction rather than a target to copy.

Then write short notes: what the song needs, what you love about the rough, what you do not, and anything that should stay exactly as it is. A few clear sentences can save a whole round of revisions.

Here is where the export lives in Logic Pro, Pro Tools, Ableton Live and FL Studio. Menus move between versions, so treat these as a map rather than a manual. The goal in every DAW is the same: one file per track, all from the same start point to the same end point, at the session’s sample rate.

Before you send anything, drop the exported files into a new, empty session and play them against your rough mix. Everything should line up from bar 1, nothing should be missing or silent, and nothing should clip. It takes a few minutes and catches most mistakes.

Zip the folder so the files travel together, then share it through a file-transfer service or a cloud storage link the engineer can download without signing up. Put your notes in the folder and in your message.

Working with Nate Lanzino? He mixes fully digital, analog or anything in between, offers mixing at Wright Way Studios in Baltimore, and works with artists in person and remotely. Call or text him to talk through your mix.