Stem Delivery Guide: What to Send Your Mixing Engineer

Stem Delivery Guide: What to Send Your Mixing Engineer

Send your mixing engineer clearly named WAV files that all start from the same point in the song, plus a rough mix, BPM, key if known, reference tracks, and notes about any effects that are part of the sound. For most full mixes, send individual tracks or multitracks rather than only grouped stems.

Bad file delivery can slow down a mix before the engineer touches an EQ. If the vocals start at different times, the beat is printed too loud, the master limiter is baked into every file, or the tracks are named "Audio 1" through "Audio 47," the first part of the job becomes cleanup and reconstruction instead of mixing.

This guide explains exactly what to send, how to name it, how to export it, when to include dry and wet versions, what not to print, and how to make the folder easy for a mixer to open. Clean delivery does not guarantee a great mix by itself, but it gives the engineer the best chance to make decisions quickly and accurately.

The Short Answer: Send Aligned WAVs and a Clear Roadmap

The safest delivery format is one audio file per important track, exported from the same start point to the same end point. If the lead vocal starts at bar one and the ad-lib starts at bar one, they line up instantly in the engineer's session even if the ad-lib does not enter until the second verse.

Along with those files, include a rough mix so the engineer can hear your intended balance, a text note with BPM and key if known, and any creative instructions that matter. Keep the note short. A mixer needs priorities, not a five-page essay.

What Your Engineer Is Trying to Avoid

Clean delivery is not about being formal. It is about avoiding mistakes that change the song before the mix starts. If an ad-lib is late by half a beat, the engineer may think it was performed that way. If a harmony is missing, the first mix pass may be built around the wrong arrangement. If the rough mix has a delay throw but the exported files do not, the mixer has to guess whether that moment was intentional.

The engineer's first job is to recreate the song correctly. Only after the files line up, the parts are labeled, and the references make sense can the creative mix begin. Good delivery makes that first stage fast and boring. Bad delivery turns it into detective work.

This is especially important when you are paying for a service with a fixed turnaround. Time spent finding missing files, repairing exports, and asking basic questions is time not spent improving vocal tone, low end, depth, impact, and translation. A clean folder gives your money more room to go toward the actual mix.

Stems vs Multitracks: Know What You Are Sending

Artists often use the word "stems" to mean any exported files. Engineers may use it more specifically. In many professional workflows, "multitracks" means individual files like kick, snare, lead vocal, double, ad-lib, bass, pad, and guitar. "Stems" can mean grouped files like drums, music, lead vocals, backgrounds, and effects.

For a full mix, individual multitracks are usually better because the engineer can control every element. Grouped stems are more common for stem mastering, live playback, remix delivery, or quick alternate versions. If you are not sure which the engineer wants, ask before exporting.

Delivery type	Example	Best use
Individual multitracks	LeadVox, DoubleL, DoubleR, Kick, Snare, Bass, Pad	Full mixing
Grouped stems	Drums, Music, Lead Vocal, BGV, FX	Stem mastering or limited revisions
Two-track beat plus vocals	Beat WAV, Lead Vocal, Doubles, Ad-libs	Vocal mix over a leased beat
Session file	Full DAW project folder	Only when the engineer uses the same DAW and plugin setup

Export Every File From the Same Start Point

This is the most important rule in the whole guide. Every file should start at the same place, usually bar one or the exact beginning of the song timeline. Do not trim the silence before a chorus harmony. Do not export the ad-lib only from the moment it starts. Silence at the beginning of a file is fine. Misalignment is not.

When every file starts together, the engineer can drag the folder into a new session and the song plays correctly. When files start at different points, the engineer has to rebuild timing manually. That can create mistakes, slow the mix down, and waste revision energy on preventable problems.

How to Handle a Two-Track Beat

Many artists do not have full beat stems. They have a stereo beat WAV from a producer, lease download, or YouTube-style purchase. That is workable, but it changes what the engineer can control. They can mix the vocals into the beat, shape the vocal pocket, clean the low mids, and make the final record translate better. They cannot individually rebalance the kick, 808, snare, melody, and hats unless those files are provided.

If you only have a two-track beat, send the highest-quality beat file you have. Do not send a low-bitrate MP3 if a WAV is available. Do not normalize it, limit it again, or export it through your vocal session with extra master effects. Send the beat clean, then send a rough mix showing where you placed the vocal against it.

Also tell the engineer if the beat is already mastered or leased with limited control. That context changes the mix approach. A crushed two-track beat may need the vocals shaped around it rather than the beat reshaped around the vocals. Clear expectations prevent unrealistic revision notes later.

Use WAV at the Session Sample Rate

Send WAV files unless your engineer asks for something else. WAV is uncompressed and widely supported across audio software. Avid describes WAV as an uncompressed export format for Pro Tools mixes, and Apple Logic Pro's export workflow lets users choose file format and bit depth when exporting tracks as audio files. The practical point is simple: do not send MP3s for mixing.

Use the sample rate of the original session. If the song was recorded at 48 kHz, export at 48 kHz. If it was recorded at 44.1 kHz, export at 44.1 kHz. Do not convert files just because you think a bigger number looks more professional. Conversion adds another place for mistakes.

Spec	Recommended delivery	Why it matters
Format	WAV	Uncompressed and easy to import
Bit depth	24-bit if available	Good headroom for mix processing
Sample rate	Match the session	Avoids unnecessary conversion
Start point	Same start for every file	Keeps the song aligned
Master limiter	Off unless requested	Preserves dynamics for the mixer
Normalization	Off unless requested	Preserves your internal balance

Do Not Print the Master Chain Into Every File

Turn off master-bus limiters, clippers, loudness maximizers, and final mastering chains before exporting files for a mix. A rough mix with those effects is useful as a reference, but the individual files should not be crushed through the master chain.

If the stems are already limited, the engineer has less room to balance, compress, EQ, and automate. Transients are already shaved. Low end may already be distorted. Vocals may already be pinned. Give the engineer clean source files and a separate rough mix that shows your intent.

Do Not Send Every Hidden Track in the Session

More files are not always better. Send every track that belongs in the song, but do not include muted experiments, old takes, unused beat versions, scratch vocals, duplicate bounces, or hidden tracks unless they matter. A folder with 120 unclear files can slow the mix down more than a focused folder with 28 useful files.

Before exporting, clean the session. Mute or remove unused tracks. Choose the final comp. Label the best doubles. Decide which harmonies are part of the arrangement. If you are unsure whether a part should stay, include it in a clearly labeled "Optional" folder or mention it in the notes.

This does not mean over-editing the song until it loses feel. It means making intentional choices. The engineer should not have to decide whether "Audio 36 copy new" is the lead vocal, an old take, or a mistake.

Send Dry and Wet Versions When the Effect Matters

Some effects are part of the production. A vocal throw, a tuned delay, a distorted ad-lib, a reverse reverb, a guitar amp sound, or a filtered vocal chop may be essential. In those cases, send the wet version and a dry version if possible.

Label them clearly:

• LeadVocal_DRY.wav
• LeadVocal_WET_reference.wav
• HookDelayThrow_WET.wav
• VocalChop_DRY.wav
• VocalChop_WET.wav

This gives the engineer options. They can keep your creative sound, rebuild it cleaner, blend it with the dry file, or use it only as a reference.

Prepare Vocal Stacks So the Lead Stays Obvious

Vocal stacks can get confusing quickly. A song may have a main lead, verse doubles, hook doubles, left and right stacks, low harmonies, high harmonies, ad-libs, whisper layers, and one-off delay throws. If those are not labeled, the engineer has to listen through everything before knowing what supports the lead.

Put the main lead first and name it clearly. Then group supporting vocals by role: doubles, harmonies, ad-libs, backgrounds, special effects. If a double should stay tucked, label it as a double. If an ad-lib should be loud for one moment, mention it in the README. If a harmony is meant to feel like a pad, say that. Good labels protect the vocal arrangement.

Also avoid combining everything into one "vocals" file unless the engineer specifically asks for a grouped stem. For a full vocal mix, separate tracks are better. The lead may need different compression than the doubles. Harmonies may need different EQ than ad-libs. If everything is printed together, those decisions are locked.

Name Files Like a Human Will Open Them

Good names save time. Bad names create mistakes. Name each file by role, section if needed, and version if needed. Avoid "final final," "new new," "Audio 12," and "bounce copy."

A clean folder might look like this:

• 01_LeadVocal_Main.wav
• 02_LeadVocal_Double_L.wav
• 03_LeadVocal_Double_R.wav
• 04_Harmony_High.wav
• 05_Harmony_Low.wav
• 06_Adlibs.wav
• 07_Kick.wav
• 08_Snare.wav
• 09_Hats.wav
• 10_808.wav
• 11_MelodyLoop.wav
• 12_Pad.wav
• 13_FX_Risers.wav
• RoughMix_DO_NOT_MIX_FROM.wav
• README_BPM_Key_Notes.txt

The numbering is optional, but it helps keep the folder readable. The role name is not optional. If the engineer has to listen to every file just to figure out what it is, the folder is not ready.

Check the Export Before You Upload It

After exporting, create a blank session and import the files. Press play from the start. This one check catches most delivery problems before the engineer sees them. If the rough mix and imported tracks do not line up, fix the export. If a harmony is missing, export it. If a file is silent, remove or replace it. If the beat is distorted, find the cleaner version.

Do not assume the zip file is correct because the export finished. Open the zip after it is created. Make sure the audio files, rough mix, and README are inside. If you are using cloud storage, confirm the link is viewable or downloadable before sending it. A broken link can delay a mix just as much as a bad export.

This final check takes a few minutes. It can save a full day of back-and-forth if the engineer starts the project at night, on a schedule, or between other bookings.

Include a Rough Mix and References

The rough mix tells the engineer what you were hearing before the handoff. It does not need to be perfect. It needs to show balance ideas, effect moments, vocal placement, and arrangement intent. Mark it clearly so nobody mistakes it for a stem.

Reference tracks are different. They show the direction you want: vocal brightness, low-end weight, width, dryness, aggression, or overall polish. Send one to three references, not ten. Too many references can create conflicting instructions.

If you are still deciding whether the service fit is right, the guide on how much mixing and mastering costs gives useful pricing and scope context before you send files.

Prepare Vocals Before Export

Do not send a messy vocal stack if you can clean the basics first. Remove obvious bad takes, mute unused tracks, label doubles and harmonies, and keep ad-libs separate from the main lead. If the vocal timing is intentionally loose, say that. If you want the engineer to tighten it, say that too.

Do not hide important decisions. If an ad-lib is supposed to be low and wide, note it. If a harmony should feel like a pad, note it. If the lead vocal should stay raw and dry, say so. Clear notes protect the song from being mixed against your intention.

What Not to Ask the Stem Folder to Explain

The folder should be clear, but it cannot replace a short creative direction. If you want the vocal darker than the rough mix, say so. If you want the beat to stay aggressive, say so. If you hate heavy reverb, say so. If the reference track is only for vocal brightness and not for low end, say that too.

Do not make the engineer guess your taste from file names alone. "LeadVocal_WET" tells them an effect exists. It does not tell them whether you love that effect, tolerate it, or want it rebuilt cleaner. The README is where you explain those priorities in plain language.

Good notes are short and ranked. For example: "Main priority: lead vocal upfront and clear. Keep hook delay throws. Do not make the master too harsh. Reference is for vocal tone, not bass level." That kind of direction is easier to act on than a long paragraph about every plugin you used.

What to Put in the README File

A small text file can prevent a long email chain. Keep it practical:

• Artist name and song title
• BPM
• Key, if known
• Session sample rate, if known
• Reference tracks
• Any files that are wet references only
• Creative notes for vocal level, effects, or low end
• What you do not want changed
• Delivery deadline, if already agreed

For more on avoiding handoff confusion, read How to Organize Files So Collaboration With Engineers Goes Faster. File organization and stem delivery are connected, but this guide focuses on the audio files themselves.

Common Stem Delivery Mistakes

• Files start at different times. This causes alignment problems.
• Only the rough mix is sent. A stereo rough mix is not enough for a full mix.
• MP3 stems are sent. MP3 is for sharing, not serious mix source files.
• Master limiter is baked into every export. This limits what the engineer can fix.
• Effects are removed with no wet reference. The engineer cannot know which delays or throws mattered.
• Files are not named clearly. The engineer has to decode the song before mixing it.
• Too many unused tracks are included. This slows decisions and creates confusion.
• No rough mix is included. The engineer cannot verify the intended arrangement.

What Clean Delivery Changes About the Mix

Clean delivery gives the engineer time to mix instead of repair. The first pass can focus on tone, balance, depth, vocal emotion, low-end control, and translation. Messy delivery shifts that time into file naming, alignment, noise cleanup, and guessing.

That also affects revisions. If the first pass is built from clean files, revisions can be creative: vocal up, snare brighter, hook wider, delay less wet. If the first pass is built from confusing files, revisions often become technical: wrong ad-lib, missing harmony, beat printed too loud, dry vocal missing, hook delay gone. Good prep protects your revision rounds.

Final Stem Delivery Checklist

1. Export one file per important track for a full mix.
2. Start every file at the same point.
3. Use WAV, ideally 24-bit if your DAW supports it.
4. Keep the original session sample rate.
5. Turn off master limiters and final loudness chains.
6. Do not normalize unless the engineer requests it.
7. Send dry and wet versions for important effects.
8. Name every file by role.
9. Include a rough mix clearly labeled as a reference.
10. Include BPM, key if known, references, and notes.
11. Zip the folder before uploading.
12. Open the zip once and confirm the files are actually inside.

Frequently Asked Questions

Should I send stems or the full session file?

Send audio files unless the engineer specifically asks for the session. Audio files avoid missing plugin, version, and DAW compatibility problems. A full session can help only when the engineer uses the same setup.

Do all stems need to start at bar one?

Yes. For a clean handoff, every exported file should start at the same point. Silence before a part enters is fine because it keeps alignment simple.

Should I leave Auto-Tune or vocal effects on?

If the effect is part of the performance or sound, include a wet version. If possible, also send a dry version. Label both clearly so the engineer knows what is creative and what is optional.

Can I send MP3 files for mixing?

No, not as source files. MP3 is compressed and throws away detail. Use WAV or the format your engineer requests. MP3 is fine for a rough reference, not for the files being mixed.

Should I remove all plugins before exporting?

Remove master-bus processing and anything you do not want locked in. Keep effects that are essential to the sound, or send both wet and dry versions so the engineer has options.

What if I only have a two-track beat?

Send the highest-quality beat file you have, plus clean vocal files, a rough mix, BPM, references, and notes. The engineer can still mix the vocals into the beat, but they will have less control over the instrumental.

The Bottom Line

Good stem delivery is not complicated. Send aligned WAV files, clear names, a rough mix, references, and a short note that explains the song. Turn off the master chain, keep creative effects documented, and make the folder easy to open. The cleaner the handoff, the faster the engineer can focus on the mix that listeners will actually hear.