You open the email expecting a Pro Tools session and a folder of raw multitracks. Instead you get eight stereo files: DRUMS, BASS, VOX, MUSIC, FX, and a couple of mystery bounces labeled “extra.” There is no session. There are no individual tracks. Welcome to mixing from stems, the situation almost every freelance engineer runs into and almost nobody plans for.
Mixing from stems is not automatically a problem. Plenty of great records get mixed this way. But it changes what you can do, what you should charge, and what you need to confirm before you accept the project. Handle it well and a stem project can be faster and cleaner than a 90-track session. Handle it badly and you spend four hours trying to un-bake a reverb that was printed onto a vocal you can no longer touch.
This post is the full breakdown. We will cover what mixing from stems actually means, why clients send stems instead of sessions, what you gain and lose, the exact questions to ask before you say yes, the stems that cause the most trouble, and how to price the work so you do not end up subsidizing someone else’s bounce settings.
What “Mixing From Stems” Actually Means
A stem is a group of related tracks bounced down to a single audio file. Instead of 14 separate drum tracks, you get one DRUMS stem. Instead of nine vocal tracks, you get one VOX stem, or maybe two if they split lead from backing. The technical definition is straightforward: a stem is a group of similar sound sources printed together as one file.
That grouping is the whole story. When you mix a full session, you have every individual element on its own fader. When mixing from stems, you have a handful of pre-combined groups and far less granular control over what is inside each one.
The number of stems varies wildly. Some clients send four (drums, bass, music, vocals). Some send twelve. The more stems you get, the more control you keep. Four stems is closer to remixing than mixing. Ten well-organized stems is genuinely workable. Knowing where a project falls on that spectrum is the first thing to figure out.
Why Clients Send Stems Instead of a Full Session
Understanding why the stems showed up helps you handle the conversation that follows. There are four common reasons, and they call for different responses.
The first is DAW mismatch. The client produced in FL Studio or Ableton and you mix in Pro Tools or Logic. Session files do not transfer between programs, so stems are the only practical handoff. This is the most legitimate reason and usually produces decent stems.
The second is a producer handoff. The beat came from a producer who sold or leased it, and the artist only ever received bounced stems, never the project. The artist literally cannot send you a session because they never had one.
The third is protecting intellectual property. Some producers deliberately withhold the full session so their sounds and chains cannot be copied. You get stems and that is all you are ever going to get.
The fourth is the one that causes trouble: the client does not know the difference. They think stems and a session are the same thing, they bounced whatever their export button gave them, and the result is a mess of inconsistent files. That client needs educating before you start.
Mixing From Stems vs. Mixing a Full Session
The core tradeoff is control. A full session gives you everything. Mixing from stems gives you groups, and you can only shape what lives inside each group as a whole.
Say the snare is too bright. In a full session, you grab the snare track and pull the top end down. When mixing from stems, the snare is glued inside a DRUMS stem with the kick, hats, and toms. Touch the high end and you affect the cymbals too. You can still get there with multiband processing and surgical EQ, but it is slower and less precise.
The same applies to balance. If the backing vocals are too loud inside a single VOX stem, you cannot simply pull them down. They are printed at that level relative to the lead. Your only options are to ask for a re-bounce or to work around it with automation and dynamic processing that treats the whole stem at once.
None of this makes the work impossible. It makes it different. The skill in stem mixing is doing more with less access, leaning on bus processing, parallel chains, and clever automation instead of per-track surgery. Engineers who are good at it often prefer the speed. Engineers who fight it are usually trying to mix stems like a full session and getting frustrated that the faders are not there.
What to Ask For Before You Accept a Stem Project
This is the part that saves you. Most stem disasters trace back to questions you should have asked at intake and did not. Before you quote or accept a project where you will be mixing from stems, confirm the following in writing.
Are the stems consolidated from the same start point? Every stem must begin at the exact same timestamp, ideally bar one, beat one, even if that stem is silent until the second chorus. If they are not time-aligned, nothing lines up when you drop them in, and you will waste an hour nudging files by ear.
What sample rate and bit depth? You want 24-bit minimum, at the session’s native sample rate. If the client bounced 16-bit MP3 stems, stop. That is not a mixable source and you need to say so before you start.
Are the stems dry or wet? Dry stems (no reverb, delay, or heavy effects printed) give you the most room to work. Wet stems with effects baked in lock you into the producer’s choices. Neither is wrong, but you need to know which you are getting so you can set expectations.
How many stems, and what is in each? Get the list before you quote. Four stems is a very different job from twelve. A VOX stem that combines lead, doubles, and ad-libs is harder to balance than three separate vocal stems.
Is there a reference or rough mix? A rough bounce tells you what the client is hearing in their head. Pair it with a clear brief. Our post on working with a client rough mix covers how to read what a reference is actually telling you.
The Stems That Cause the Most Trouble
Some stems are perfectly clean. Others are landmines. Knowing the difference before you commit protects your time and your quote.
Printed effects are the most common headache. A vocal stem with a long reverb tail baked in cannot be made drier. If the client later asks for a tighter, more intimate vocal, you physically cannot deliver it from that stem. Flag baked-in effects at intake and confirm the client is okay living with them.
Glued bus processing is the next one. If the producer slammed a limiter across the drum bus before bouncing, the dynamics are already crushed. You inherit that crushed sound and cannot recover the transients. The drums will hit how they hit, and no amount of clever processing brings back what the limiter ate.
Mismatched lengths are a pure time-waster. One stem starts at bar one, another starts at the first downbeat of the verse, a third has two bars of count-in. Aligning these by hand is tedious and error-prone. This is exactly why you ask about consolidation up front.
Pre-chopped vocal comps are sneaky. Sometimes a VOX stem is actually a rough comp with audible edit clicks, breaths cut too tight, or two takes crossfading badly. You can clean some of it, but you are now doing editing work that was never in your quote. This is the same trap as the one we covered in the fix it in the mix post, just wearing a different outfit.
How to Price Mixing From Stems
Pricing is where engineers lose money on stem projects, because stems can go either way. Clean, well-organized stems are often faster to mix than a sprawling session, which argues for your standard rate or even a slight discount. Messy stems with baked-in problems are slower and more frustrating than a full session, which argues for a premium.
The answer is to never quote a stem project blind. Look at the files first. Ask for the stems before you give a final number, or quote a range and confirm it once you have heard what you are working with.
Build the assessment into your process. When the stems arrive, spend ten minutes auditioning them before you lock a price. Are they dry or wet? Consolidated or scattered? Clean or already mangled? That ten minutes is the difference between a fair quote and a project you resent.
If the stems are a mess, you have the same three options you would offer for any bad source material. Ask for a clean re-bounce. Quote the extra cleanup time as a separate line. Or accept the limitation in writing and mix what you have. Whatever you decide, put the scope in your agreement. The mix engineer contract post covers how to write a clause that distinguishes a standard stem mix from a rescue job, so you are not absorbing editing work you never priced.
When to Ask for the Full Session Instead
Sometimes the right move is to push back and request the actual project files. Not always, because plenty of clients genuinely cannot provide them, but often enough that it should be a reflex.
Ask for the full session when the stems are too baked-in to deliver what the client is describing. If they want a dramatically different vocal sound and the vocal stem has effects printed on it, the session is the only path. If they want surgical changes to one element inside a glued stem, the session is the only path.
Ask when the stems arrive broken. Mismatched lengths, wrong sample rate, MP3 sources. Rather than build a quote around fixing all of that, request a proper bounce or the session itself.
But do not demand the session reflexively. If the client mixes in a different DAW or only ever had stems from a producer, the session does not exist and pressing for it makes you look like you do not understand how modern projects come together. Read the situation. Mixing from stems is a normal, professional way to work, not a downgrade you should resist on principle.
Set Up the Stem Handoff So Files Arrive Clean
The cleanest stem project is the one where you defined the handoff before a single file moved. Most stem chaos comes from clients guessing at what you need and exporting whatever their DAW gives them by default.
Tell clients exactly how to send stems, in plain language they can follow. Consolidate every stem from bar one. Bounce 24-bit WAV at the session sample rate. Send dry stems plus a rough mix for reference. Label each file clearly. Five minutes of instruction prevents most of the problems in this post.
The handoff itself matters too. When clients send tracks through a structured upload like session.trackbloom.com, the stems arrive grouped by instrument instead of dumped into a zip with no order. You see what is there before you commit, you can flag a baked-in effect or a wrong sample rate at the door, and the project starts with a defined shape instead of a guessing game. Compare that to the usual WeTransfer link where files land in a random pile and the problems only surface once you have already loaded the session.
A clean handoff does not turn bad stems into good ones. It does mean you catch the issues before you have sunk three hours into a mix, which is most of the battle.
Make Stem Projects Work For You, Not Against You
Mixing from stems is a permanent part of the job now. More clients produce in their own DAWs, more beats change hands as bounces, and more artists will only ever have stems to give you. Fighting that reality is a losing game.
The engineers who thrive with stems are the ones who treat them as a known format with known constraints, not as a broken version of a real session. They ask the right questions at intake. They audition before they quote. They price the messy ones as the cleanup jobs they are. And they educate clients on how to bounce a clean set of stems so the next project lands easier than the last.
Do that, and a folder of stems stops being a surprise you dread and becomes just another way good work comes in the door.