Finishing and Exporting: Bounces, Stems, and Release-Ready Files
The unglamorous last mile of a track: bouncing stems, choosing sample rate and bit depth, checking your mix on every system, and printing files that are ready to release.
The Last Mile Nobody Talks About
The song is finished in your head long before it is finished on your drive. Exporting is the unglamorous stretch between a mix you love in your session and a file a stranger can play on a phone in another country, and it gets skipped in tutorials because none of it is creative. It is also where good tracks quietly go wrong: a bounce that clips on the last chorus, a file printed at the wrong sample rate, a master that sounds thin on earbuds because you only ever checked it on your monitors.
None of this is hard. It is a checklist you run the same way every time, and once it becomes a habit you stop thinking about it at all. This guide walks the last mile in order — bounce your stems, lock your sample rate and bit depth, quality-check on systems that are not your studio, and print files that are ready to upload — so the version the world hears is the version you actually made.
Bounce Your Stems Before Anything Else
A stem is a bounce of one part of your mix — drums, bass, vocals, the synths — printed as its own full-length audio file, with every stem starting at the exact same point so they line back up when you stack them together. Bounce stems before you call a project finished, even if you are sure you will never need them. A mastering engineer may ask for them, a remixer may want them, a sync placement may require a separate instrumental and a-cappella, and future-you will be grateful the day a plugin you leaned on no longer opens.
The rules are simple and worth following exactly. Start every stem at bar one so they drop in perfectly aligned, let effect tails ring out instead of chopping them off, and print with your master-bus processing bypassed so each stem is clean and the set recombines back to your mix. Name them clearly and consistently — kick, bass, lead_vocal, backing_vox — because a folder of files called audio_1 through audio_12 helps nobody. If your track has vocals, always print a clean lead-vocal stem and a full instrumental; those two are the files people ask for most.
Sample Rate and Bit Depth: Decide Once, Never Upsample
Sample rate and bit depth describe how your audio was captured as numbers: the sample rate is how many snapshots are taken per second, and the bit depth is how finely each snapshot is measured. Two practical rules cover almost every situation. Work and deliver at the sample rate your project was created in — commonly 44.1 or 48 kHz — and never upsample a finished file to a bigger number hoping it sounds better. Apple, in its Apple Digital Masters guidelines, tells engineers to use 24-bit sources and send the highest-resolution master file possible, and warns plainly that upsampling will not recover or add anything that was not already in the original recording [1].
Bit depth is where you make one real decision. Mix and master at 24-bit so you have headroom and detail to work with, and keep that 24-bit file as your archive master. When a destination specifically needs 16-bit — the old CD-quality baseline that plenty of stores still deliver — reduce to it as the very last step and turn on dither, a tiny layer of shaped noise that masks the rounding errors you would otherwise hear as faint crackle in quiet fades. Dither once, at that final bit-depth reduction, and never again.
Quality-Check on More Than Your Monitors
Your studio monitors are the one place your track is guaranteed to sound fine, because you mixed it there. The entire point of quality-checking is to hear it everywhere else: the phone speaker, cheap earbuds, a laptop, the car, a Bluetooth speaker across the kitchen. Each system exposes a different weakness — the phone tells you whether the vocal and melody survive with almost no low end, the car tells you whether the bass is boomy, the earbuds tell you whether the top end is harsh. If the track holds up across all of them, it will hold up out in the world.
Two checks catch the most embarrassing mistakes. Collapse the mix to mono and confirm nothing important disappears, because a wide, stereo-heavy sound can partially cancel when a club rig or a phone sums to mono, and a lead vocal that vanishes in mono is a genuine problem. Then play the exported file itself, start to finish, and actually listen: a bounce can introduce a click at the edges, a clipped peak in the loudest section, or a fade that cuts a beat too soon. Check the printed file, not only the session, because the file is what ships.
Print Release-Ready Files
For distribution, print one master as a WAV at your project's sample rate and 24-bit depth, and resist the urge to crush it loud first. Every major streaming platform normalizes playback toward a target loudness, so a file mastered painfully loud simply gets turned back down on the listener's end, and all you are left with is the squashed dynamics you traded away. Spotify, for example, normalizes tracks to -14 LUFS and recommends keeping your true peak below -1 dBTP so lossy encoding does not push transients into distortion [2]. Leave that peak headroom and master for balance, not for a number on a meter.
The rest is bookkeeping that keeps the release clean. Give the file a clear name, fill in the metadata your distributor asks for — track title, artist, and any contributors — and upload a lossless file such as WAV or FLAC rather than an MP3, so the store encodes from full quality instead of from something already degraded. Then keep your own archive: the 24-bit master, the stems, and the project file, stored somewhere you will still have them in a year. The uploaded file is the destination, but the archive is what lets you cut an instrumental, fix a typo in the credits, or remaster later without starting from scratch.
Closing the Idea-to-Release Loop
Finishing and exporting is the far end of a loop that started with a sound you loved. Somewhere back at the beginning you heard a bass, a vocal chain, or a lead you wanted, and you reverse-engineered it into your own track — that is the whole SoundArc idea: feed it a sound you love and get back a real, editable preset instead of guessing at parameters for an hour. Everything between that moment and this one was you turning a starting point into a song.
The last mile is what makes it real. A preset in your session is potential; a checked, correctly exported, release-ready file is a record other people can actually play. Print your stems, lock your sample rate and bit depth, quality-check on systems that are not your studio, and export a clean master, and you have closed the loop from a sound you admired to a finished track of your own — vocals included, which is exactly the corner VocalArc is built for. Reverse-engineer what you love, then carry it all the way to a file.