Final delivery closes one production schedule. It does not necessarily close the life of the music. Months later, a producer may need an approved master, a shorter end credit or an alternate for a presentation. The request sounds small until the team discovers several folders full of nearly identical exports.
I prefer to plan the archive while the score is being made. Its purpose is to recover an approved state, understand which picture it belongs to and identify what can be changed. Storage capacity is a poor measure of success. A useful archive lets somebody outside the original session find the right material with confidence.
Start with the future request
Before discussing source sessions, I ask production what it expects to do later. Relivering an identical master needs a different package from replacing a soloist, changing orchestration or rebuilding the music mix. Asking for everything can produce a large folder without solving any of those tasks.
I separate preservation, adaptation and new creative work. Preservation means recovering the approved result. Adaptation uses prepared elements such as stems and alternates. New creative work may require the composer, a session rebuild, fresh programming or another recording. Those are different levels of access and effort.
This belongs in the conversation when hiring a composer for hire, alongside the post schedule and deliverables. A complete source session can be difficult to use on another system. A smaller, documented set of masters may answer a simple request immediately. The package should reflect the expected use.
Keep the approval attached to picture
The archive starts with a reference. I retain the picture filename, its version, start timecode and the cue identifiers associated with approval. Where the agreed package includes it, a reference movie helps somebody recover the entrances and exits without rebuilding the entire editorial timeline.
Picture lock is useful context, but the record should still identify the actual delivered cut. A late change can affect a placement after the score was approved. A music editor may have shortened a cue during the dub. The archive should make those relationships visible.
I also label exceptions. A festival version may precede final recording. A mockup may have been deliberately retained in one scene. An approval can apply to the composition while a technical change remains open. The archive should preserve that meaning instead of reducing every decision to the word final.
Separate listening files from working material
I keep stereo masters, stems, alternates and documentation in clearly identified areas. Source sessions belong in a separate package. Editorial should be able to retrieve the delivered music without opening the composition workspace.
Filenames carry the project, cue identifier, revision and file type. A short convention used consistently is more useful than a long convention used unpredictably. A reference table links those identifiers to cue titles and approved versions.
Technical details remain explicit. Sample rate, bit depth, common start and duration should not depend on somebody guessing what the original dub stage expected. A brief readme can record the delivery specification and any relevant exceptions. That document helps the next team distinguish an intentional difference from a damaged or incorrect export.
Preserve the logic of the stems
Stems need to remain connected to the mix they represent. I check their start, duration and contents. I document shared processing when it explains why summing the stems may differ from the main master.
The producer’s practical question is what the files allow a future mixer to do. Removing a melody requires an appropriate separation. Reducing a pulse requires access to that element. A package split by instrument families may support some changes and make others difficult.
I describe the available flexibility instead of promising unlimited reconstruction. Stems support editorial and mix changes. They do not necessarily expose individual notes, recreate virtual instruments or preserve the full processing chain. Knowing that boundary makes a future request easier to scope.
A session file needs an environment
The DAW document points to audio, instruments and processing. Reopening it may require compatible software, particular versions, sound libraries and the necessary licences. Keeping the document alone does not preserve that environment.
When session archiving is in scope, I include a dependency list and gather the associated media. Depending on the expected reprise work, I also render selected virtual instruments or processed parts to audio. Those renders preserve an audible result when a particular tool is unavailable later.
I would not promise effortless reopening years into the future. I can reduce unknowns and distinguish editable material from audio that has been fixed. Production can then decide whether that level of preservation matches the project’s likely needs and available resources.
Test recovery before everyone moves on
An archive needs a recovery check. I verify that the expected files are present, open correctly and match the reference delivery. For a source session, a reopening test can expose missing media or overlooked dependencies while the original team is still available.
I recommend that production maintain copies in separate locations and identify the person responsible for checking them. The exact arrangement depends on its infrastructure. The aim is to avoid relying on one machine or a former collaborator’s personal account as the sole route to the score.
The test should fit the package. Checking a master folder is a different task from validating a large orchestral session with many dependencies. That work needs a place in the delivery scope and calendar. It becomes harder to schedule once the production has dispersed.
Make later requests traceable
The archive note identifies the package version, the available material and the appropriate contact for reprise work. It can also point production back to the agreed limits on use and modification. Possession of source files does not itself establish the permitted uses; the relevant project agreements need to be checked.
I ask that subsequent versions join the history instead of overwriting the original delivery. A new cutdown receives a new identifier. A rebuilt mix retains its relationship to the approved master. The next request then starts from a visible chain of decisions.
For an initial conversation, send the possible future uses, expected deliverables and the contact who will maintain the archive. I can propose a package that supports those needs and include the preparation in the score scope. You can review my work and send the project details.