How to Prepare and Export Stems for Online Mixing

Sending clean, clearly labelled audio files is one of the easiest ways to get a better online mix. It reduces guesswork, prevents avoidable technical problems and lets the engineer spend more time improving the record.

The word stems is often used for every exported file, but there is a useful distinction. Multitracks are the individual recorded or programmed parts: kick, snare, bass, lead vocal and so on. Stems are grouped submixes, such as all drums in one stereo file or all backing vocals together. Unless your engineer specifically requests stems, individual multitracks normally provide the most control.

Here is a practical process for preparing either format.

1. Confirm the arrangement before exporting

Open the final production session and listen from beginning to end. Check that the song structure is approved, the preferred vocal takes are active and no experimental tracks are playing by accident. Remove obvious false starts, muted ideas and unused playlists from the delivery copy of the session.

Do not continue rewriting the production after you send the files. A new chorus length or replacement vocal changes the mixing job and can make earlier work unusable. If a creative decision is still open, tell the engineer before the session begins.

Leave a short amount of silence before the first musical event and let delays or reverbs decay naturally after the ending. Do not cut the export exactly at the final beat.

2. Print sounds that define the production

If a sound depends on a software instrument, amp simulator, special effect or unusual routing, print an audio version that captures it correctly. A mixer may not own the same plugin, preset or sample library.

For important processed parts, it can help to provide both versions:

• `Lead Vocal PRINT` with the creative effects you consider essential.
• `Lead Vocal DRY` without those effects, where appropriate.
• `Guitar AMP` with the chosen amplifier tone.
• `Guitar DI` if you genuinely want the option to re-amp it.

Do not remove processing simply because someone said all files must be dry. Timing effects, distortion, filter moves and sound-design chains may be part of the arrangement. Keep anything that defines the identity of the record and explain it in your notes.

3. Clean the session without damaging natural performances

Remove accidental clicks, count-ins that should not be heard and unwanted noise between sections. Add short fades to edited clips so that cuts do not create clicks. Check vocal comp boundaries and crossfades at a normal listening level and on headphones.

Be careful with aggressive noise reduction or hard gates. A gate that clips the first consonant of a word is harder to repair than a little room noise. Natural breaths can support phrasing; remove only the ones you deliberately do not want.

Tune and time performances to the degree you want before mixing, unless editing is part of the booked service. If you need help with vocal tuning, read our guide to professional vocal tuning and review the available vocal-production service.

4. Use one common start point

Every exported file should begin at the same timeline position, even when that instrument enters late. This keeps all parts in sync when they are imported into a new session.

For example, if the song starts at bar 3, export every track from bar 1 or another agreed position. A shaker that first appears in the final chorus should still include silence from the common start point.

Before sending anything, create a blank test session and import all of the exports at the same position. Press play. If the arrangement rebuilds correctly without manual nudging, the timing is ready.

5. Choose sensible audio settings

Export WAV or AIFF files at the session's native sample rate and bit depth. If you recorded at 48 kHz and 24-bit, keep that format. Upsampling a 44.1 kHz session does not add missing detail, and converting sample rates unnecessarily creates another stage where mistakes can occur.

Use mono files for genuinely mono sources when your software handles them reliably; use stereo for stereo instruments, stereo microphones and effects returns. Do not convert everything to stereo merely to make the files look consistent.

Avoid MP3 or other lossy formats for mix sources. They discard information and can produce audible problems after further processing.

6. Preserve headroom and avoid accidental clipping

Individual files should not clip. Watch both the track and bus routing while exporting. If your production is hitting the stereo output too hard, solve that in the session rather than simply pulling down a clipped bounce.

There is no requirement to make every file peak at the same number. Healthy, unclipped levels are more important than chasing an arbitrary target. Do not normalise each export unless the engineer has specifically requested it; normalisation changes the level relationships and offers no mixing advantage.

Provide your current stereo rough mix as a separate reference. It shows the balance, effects and energy you have been hearing. If you used a limiter only to make that rough mix louder, keep it on the reference but leave it off the individual multitracks.

7. Label files so another person can understand them

Use short, descriptive names. A practical pattern is:

`SongTitle_Instrument_Detail`

Examples include `NightDrive_Kick`, `NightDrive_BassSynth`, `NightDrive_LeadVox` and `NightDrive_BGV_High_L`. Avoid names such as `Audio 34`, `final2new` or a row of duplicate `Lead Vocal` files.

Numbering broad groups can also keep imports organised:

• `01 Kick`
• `02 Snare`
• `10 Bass`
• `20 Guitars`
• `30 Lead Vocal`
• `40 Backing Vocals`

Make it obvious when a file is mono, stereo, dry, printed or an alternate. The engineer should not have to audition six mystery files to discover which one is approved.

8. Include a concise project note

Your note should contain the song title, tempo, time signature, sample rate, approved rough mix and a few creative priorities. Mention references for direction, not as instructions to copy another record.

List any effects that must remain, any edit you are unsure about and the two or three issues you most want the mix to solve. Include lyric sheets when detailed vocal automation or explicit-language checks matter.

Keep feedback in one document. Conflicting messages across email, chat and voice notes slow down revisions.

Final delivery checklist

Before uploading, confirm that:

• every file starts at the same point and plays in sync;
• the sample rate and bit depth match the source session;
• there are no clipped exports, missing files or cut-off effect tails;
• tracks have clear, unique names;
• the approved rough mix and notes are included;
• the folder contains only the final delivery, not old versions; and
• you have listened to the exported files rather than assuming the bounce worked.

Zip the organised folder so that filenames and subfolders remain together during transfer.

Frequently asked questions

Should I send stems or individual tracks for mixing?

Send individual multitracks unless the engineer asks for grouped stems. Multitracks allow more detailed control; stems can be useful when a complex sound is already balanced or the original session cannot be reconstructed.

Should I remove all plugins before exporting?

No. Print creative processing that defines the sound. Remove unnecessary master-bus loudness processing from the mix files, and provide dry alternatives only where they are genuinely useful.

Can I send my DAW session instead?

Only when the engineer has confirmed compatibility. Consolidated WAV files are more reliable because they do not depend on the same software, plugins or sample libraries.

Ready to have your song mixed?

Explore our professional online mixing service, hear the before-and-after examples, or contact Mix & Master Engineers if you want us to review your delivery before the session begins.