How Many Alt Mixes and Stems Should You Deliver for Each Song?

As many as were agreed, and not one more. That is the honest answer, and the useful version of it is not a number but a short list of versions that earn their place by who actually uses them, settled before anything renders.

The Answer Is a List, Not a Number

The question usually arrives from one of two anxieties. Either the engineer wants to look thorough and is tempted to send everything a session can print, or the client said "send everything" and nobody knows what "everything" means. Both resolve the same way: each version of a mix exists for a specific downstream user, so the delivery list is built by asking who will use each file, not by picking a quantity. A version with no user is weight in the folder; a version with a user who never got it is a stalled deal three weeks later.

Oversending is not generous, either. Nine unrequested alternates hand the client a decision they did not ask for, imply a scope nobody priced, and bury the file that matters in a folder of lookalikes. The mix delivery checklist covers the whole handoff; this piece settles its most negotiated line.

Why the Number Keeps Creeping Up

Version lists ratchet. A one-off request, the acapella a remixer needed once, quietly becomes tomorrow's default because it was delivered once and nobody wrote down that it was an exception. Three projects later the default delivery has grown to a dozen files per song, nobody remembers deciding that, and the rendering time has scaled with it. The ratchet only turns one way because saying yes to one more version is always locally cheap; what makes it expensive is that versions, once delivered, become expected.

The defense is not refusing requests. It is keeping a written baseline, so every addition is visibly an addition, priced or at least acknowledged, instead of silently joining the default.

The Versions That Earn Their Place

Three rows deserve the longer look. The instrumental is the version to render even when nobody asked, because it is the one most often requested after the fact and it costs nothing to print alongside the mix. The TV mix is the opposite: specific enough (lead vocal out, backing vocals in) that rendering it speculatively only makes sense when live performance is plausibly in the song's future. And the acapella deserves a conversation before it exists at all, because it is the version that travels furthest from your control once it leaves.

The last two rows are briefs, not toggles. A clean version is an editorial decision made with the artist, which lyrics are replaced, reversed or muted, not a technical pass an engineer should improvise. A radio edit is structural: it needs a defined start and end, usually a sign-off on where the cut lands, and it is the version where two markers in the session become the specification.

Stems Are a Different Question

Stems are a multiplier, not one more row on the version list. One song delivered as eight category stems in two formats is sixteen files before a single alternate version joins them, and a five-song EP turns that into eighty, which is why stems are agreed, never assumed. The mechanics of rendering them repeatably are covered in our guide to stem export automation in Pro Tools; the scoping question comes first. The agreement has four parts: whether stems are wanted at all, which groups count as a stem, whether they are wet or dry, and whether the set must sum back to the mix. Confirm, separately, that stems means grouped submixes and not multitracks, because clients use the words interchangeably and the two differ by an order of magnitude in file count.

Every stem in the set follows the same two technical rules, whatever was agreed about the rest. A common start point, usually the session start, so every file lines up on sample one in whatever DAW receives it, with nobody nudging regions by eye. And effects tails rendered complete: a stem clipped where the audio stops loses its reverb into the silence, and the receiving session hears the join every time that group re-enters.

Agree the List Before Anything Renders

The cheap moment to settle the list is at booking, in writing, as part of scope. One line does it: delivery includes the main mix, instrumental and MP3 copies; TV mix, acapella, edits and stems on request as additional deliverables. From then on the list is a menu, not a surprise, and the question "how many?" stops being asked per song. Repeat clients get the same list reused, which is also how the list becomes a studio standard instead of a negotiation.

When a request lands mid-project anyway (the late TV mix, the suddenly needed acapella), the list absorbs it as one added line item rather than a reopened scope, because the baseline was written down.

A baseline worth stealing has three tiers. Always: the main mix, the instrumental and the MP3 copies, rendered on every song without discussion. On request: the TV mix, the acapella, the clean version and the edits, each one a named line item when asked for. Scoped separately: stems, because they multiply files and responsibility in a way single versions do not. Three tiers fit in three sentences of a booking email, and they turn every future "how many?" conversation into a pointer at a document.

Where the List Stops Costing Anything

Once the list exists, rendering it is the part that automation genuinely removes. In fMusic Suite each version is a print carrying its own recipe: the instrumental and the TV mix as custom prints, the radio edit as a print whose range is the span between two markers, stems as category prints that select their tracks through the session's structure, and the MP3 copies as one extra line in each print's format list rather than a second render. The whole model is in Every Print Its Own Recipe, and the full run on one song, spec to folder, is Delivery Day. Prints saved as favorites mean the agreed list renders identically on every song of the project, previewed before anything bounces, in one pass.

What fMusic does not do is decide the list. Which versions the project needs, what stems mean for this client and what was priced stay in the agreement; the tool renders the agreement exactly, which is precisely why the agreement has to exist.

Frequently Asked Questions

How many versions of a mix should I deliver by default?

The main mix, an instrumental and lightweight MP3 copies cover most deliveries. Everything else, TV mix, acapella, clean version, edits and stems, earns its place by an actual downstream user and belongs in the agreed scope.

What is a TV mix and when is it needed?

A version with the lead vocal muted and backing vocals kept, used for televised and live performances where the artist sings over the track. Needed when performances are part of the song's future, not as a default.

Should I send stems with every song?

No. Stems multiply the file count and the scope, so they are delivered when agreed, in the grouping that was agreed, with a common start point and complete tails.

What is the difference between stems and multitracks?

Stems are grouped submixes, one file per group such as drums or vocals. Multitracks are every individual track in the session. Clients say stems for both, so confirm which is meant before rendering either.

Should I charge for alternate versions?

Scope them, whether or not you price them separately. A written baseline list with extras on request keeps late additions as line items instead of renegotiations. This is a scoping practice, not legal or pricing advice.

What is the fastest way to render an agreed list of versions?

As saved prints: each version carries its own source, range and formats, stems come from categories, and the whole list renders in one pass after a preview. That is the delivery model in fMusic Suite.

The number of alt mixes and stems is whatever the written agreement says, built version by version from who will use each file. fMusic Suite renders that list as prints in one pass, and the export model is at forte-ai.com/fmusic.