How to Prepare Stems for a Mixer

Sending stems to a mixer is one of those tasks that looks simple until you’re on the receiving end of a badly prepared handoff. Wrong sample rate. Random file naming. A stereo “vocals” stem that’s actually seventeen different lead vocal takes smashed together at 0 dBFS. A mono voiceover exported as a dual mono stereo file, using twice the space and confusing everyone.

Getting stems right is a skill. Here’s how to do it.

What stems are and what mixers actually need

Stems are grouped exports of your session tracks — subgroups of related audio that give a mixer control over the major elements of your project without handing them a hundred individual tracks.

What constitutes a stem depends on the project. For a song, typical stems might include: drums, bass, guitars, keys, lead vocals, backing vocals, and FX/atmospheres. For a film score: strings, brass, woodwinds, percussion, synths. For a podcast or documentary: narration, interviews, music, and SFX.

The key is that each stem should contain tracks that you’d want the mixer to treat together. Drums and bass have very different dynamics and frequency relationships — they’re separate stems. Lead vocal and background vocals might be separate stems if the mixer needs to balance them against each other, or a single vocal stem if they’re always moving together.

What mixers don’t want: stems that are already mixed and limited. Your master bus processing should be bypassed when exporting stems. The mixer is doing the mix — they need raw material with headroom, not a pre-mixed file they can only adjust with a fader.

How to group and bus your tracks

Before you export anything, your routing has to be right.

Send all related tracks to a dedicated bus channel — drums to a drum bus, vocals to a vocal bus, and so on. That bus is what you’ll export. This means the bus itself should have only the processing that belongs to the group (bus compression, EQ that shapes the group as a whole). Individual track processing — the compressor on the snare, the saturation on the bass — stays on those tracks and prints into the stem.

Print any effects you want the mixer to hear in the stem. Reverb tails, delays, modulation — if it’s part of the sound you’re delivering, it either needs to be printed on the stem or sent as a separate FX return stem. Don’t leave it unprinted and assume the mixer will recreate it.

One common mistake: sending a “drums” stem that has all the drum channels routed individually, with no actual sum. The mixer then has to figure out how to recombine them. If you’re sending a stem, send a stem — one file that represents the fully printed group.

Give yourself headroom. Stems should peak around -6 to -3 dBFS, not 0. The mixer needs room to balance everything against each other. If your drum stem is slamming at 0, you’re starting from a compromised position.

Export settings: format, sample rate, bit depth

Match the session’s sample rate and bit depth. Don’t convert. If you worked at 48kHz/24-bit, export at 48kHz/24-bit. The mixer will open these files in their session — you want a clean format match, not a conversion artifact introduced at the handoff stage.

Format: WAV is the safest choice for stems. It’s universally supported, supports 24-bit and 32-bit, and handles broadcast metadata (BWAV chunks). AIFF is also fine for most professional contexts. Avoid MP3, AAC, or any lossy format for stems — you’re handing off working material, not a consumer file.

Sample rate: 48kHz for anything that will sync to picture. 44.1kHz for music-only projects where that’s the session rate. Check what the mixer expects if you’re unsure. Sending them files that need conversion on their end is an unnecessary friction point.

Bit depth: 24-bit minimum. 32-bit float is acceptable and gives the mixer extra headroom to catch any hot transients, but confirm they can receive it. 16-bit is not appropriate for stem handoffs — save that for consumer delivery only.

One more thing: export all stems starting from bar 1 / the absolute start of the session. Every stem should be exactly the same length, even if some tracks don’t have audio for the first 30 seconds. This is how the mixer places everything in sync without manually aligning 18 files.

Making sure your mono tracks are actually mono

This is a step most people skip, and it causes real problems.

A lot of files — especially voiceover recordings, room mics, spot effects — come in as stereo files even when they’re carrying a mono signal. Identical content on both channels. You place them on a stereo track in your DAW and they look and behave like stereo files. But they’re not. They’re dual mono: a single signal padded to two channels.

When you print a stem containing those tracks, the mono signal gets exported as stereo. Your mixer receives a stereo stem file where left and right are identical. Every stereo plugin they put on that track processes two identical channels for no reason, using double the CPU. And the file is twice as large as it needs to be.

Worse, if they’re expecting a mono voiceover stem and they receive a stereo file, it can cause unexpected panning behavior or metadata conflicts in their session.

Before you print stems, go through your tracks and verify that anything recorded as a mono source is actually set to mono in your session. For individual files, this means checking whether L and R are truly different signals or not.

For large projects — or if you’ve received audio from multiple sources and aren’t sure of the provenance — checking this manually across dozens of files is tedious. MonoMate is a macOS and Windows app that automates this: drag in a folder of files, it analyzes every one, identifies the dual mono cases, and converts them to true mono. Saves real time, especially if you’re dealing with large SFX pulls or multi-source dialog that came in from different vendors.

Get your channel format right before you export. The stem you send should reflect the actual content.

Naming your stems so the mixer doesn’t hate you

File names communicate everything when you’re not in the room.

A good naming structure for stems:

ProjectName_StemName_v1_48k24b.wav

For example:

UrbanFilm_Drums_v2_48k24b.wav
UrbanFilm_LeadVocals_v2_48k24b.wav
UrbanFilm_SFX_v2_48k24b.wav

A few non-negotiable rules:

No spaces. Spaces break scripts, confuse certain DAWs, and make terminal/command-line work harder. Use underscores.

Include the version. When you inevitably send a revised stem, the mixer needs to know which file supersedes which. _v1, _v2, _RevisionB — whatever system you use, be consistent.

Describe what’s in it, not where it came from. BusGroup3.wav means nothing. BackingVocals.wav means something.

Be consistent across all files. If your drum stem is ProjectName_Drums_v1.wav, your bass stem shouldn’t be BASS-STEM-FINAL.wav. Pick a convention and apply it to every file in the handoff.

Include the technical specs in the name or in a separate text file. The mixer shouldn’t have to open the file to find out what sample rate it is.

What to include in your handoff notes

A stem handoff without notes is an incomplete handoff. Write a short document — even a plain text file — that travels with the stems. It should cover:

Session info: Project name, tempo (or tempo map if it changes), key, total duration, sample rate, bit depth.

Stem manifest: A list of every stem file and what it contains. Drums = kick, snare, overheads, room. Vocals = lead vocal only (backing vocals are a separate stem). This prevents the mixer from guessing.

Technical notes: Any unusual routing decisions. Any tracks where you’ve left processing unprinted and why. Any files that are intentionally mono vs. stereo.

Reference track: If you have a rough mix or a reference for the target sound, include it. The mixer doesn’t need to guess where you’re headed.

Known issues: If there’s a click in bar 32 of the guitar stem that you couldn’t clean up, say so. If the room mic has a hum in the low end, flag it. Better to warn them than have them spend 45 minutes diagnosing it.

This document doesn’t need to be long. Half a page is enough. It just needs to exist.

Checklist

Before you send your stems: