---
title: "Archiving workflow: long-term preservation"
url: https://postproduction.studio/en/workflows/archiving-workflow
language: en
description: "Master the archiving workflow in post-production. Learn about LTO, LTFS, storage strategies, checksums, deep archiving, and metadata cataloging."
updated: 2026-10-25
---

# Archiving workflow: long-term preservation

Level:Advanced

Section: Workflows

14 min

## TL;DR · Key takeaways

* •Long-term archiving preserves a production's assets for future re-mastering, restoration, or distribution: camera RAW, project files, masters, deliverables, paperwork — all stored under controlled conditions.
* •LTO tape standard: LTO-9 (18 TB native, 45 TB compressed, ~30-year archival life), LTFS file system, two-copy minimum on geographically separated tapes, vendor-neutral readers (Quantum, IBM, HPE).
* •What to archive: camera-original RAW (essential for re-grading), color-graded masters (per delivery format), VFX project files (Nuke/Houdini scripts, render assets), sound stems and Pro Tools sessions, editorial AAF.
* •Cloud archival alternative: AWS S3 Glacier Deep Archive (~$1/TB/month), Backblaze B2 (~$5/TB/month), Wasabi (~$6/TB/month), Google Coldline — cheaper but ongoing cost vs LTO's one-time hardware buy.
* •Best practices: 3-2-1 rule (3 copies, 2 media types, 1 offsite), checksum verification on creation and periodic re-verification, MD5 or xxHash for all files, metadata documentation (DIT logs, EDL, technical spec sheets).

The archiving workflow is the final, yet arguably the most enduring phase of the entire post-production pipeline. Long after the wrap party is over, the theatrical run has ended, and the streaming numbers have peaked, the archive is what remains. Archiving is fundamentally different from a simple "backup." A backup is a short-term failsafe designed to protect active data against sudden hardware failure or accidental deletion during active production. An archive, conversely, is a highly structured, meticulously cataloged repository designed for long-term preservation and retrieval, ensuring that the original assets, the final masters, and all intermediate elements survive for decades. In the era of celluloid film, archiving meant storing physical negative reels in temperature-controlled salt mines. Today, in the era of digital cinematography, where a single feature film can generate hundreds of terabytes of unstructured raw data, archiving is an immensely complex IT challenge. Digital data is inherently fragile. Hard drives suffer from mechanical failure and "bit rot," proprietary software formats become obsolete, and institutional memory regarding folder structures vanishes when personnel move on. Therefore, the modern archiving workflow relies on agnostic open-source file systems, rigorous checksum verification, physical tape media (LTO), and comprehensive metadata cataloging. The goal of this phase is not just to save the data, but to guarantee that a technician thirty years from now can open the archive, understand its structure without external guidance, and perfectly restore the film to its original glory. This comprehensive guide details the architecture of modern post-production archiving, exploring deep storage strategies, long-term format considerations, the critical distinction between master deliverables and archival masters, and the best practices required to future-proof digital cinema.

## Table of Contents

* [1. Definition and challenges of archiving](#section-1)
* [2. Archiving levels (working, vault, deep)](#section-2)
* [3. Long-term archiving formats](#section-3)
* [4. Metadata and cataloging](#section-4)
* [5. Storage: LTO, cloud, and disks](#section-5)
* [6. Integrity verification and migration](#section-6)
* [7. Master archive vs deliverables](#section-7)
* [8. Documentation and traceability](#section-8)
* [9. Best practices and common pitfalls](#section-9)

## 1. Definition and challenges of archiving

Archiving in post-production is the systematic process of securing, cataloging, and storing all digital assets associated with a completed project to guarantee their long-term survival and future accessibility. The primary challenge of digital archiving is technological obsolescence. Unlike physical film stock, which can theoretically be held up to a light source and understood by the human eye even centuries later, digital files require specific hardware, operating systems, and decoding software to be interpreted. If any link in that interpretive chain breaks, the data becomes an unreadable sequence of binary code.

Another massive challenge is the sheer volume of modern digital production. A 4K or 8K multi-camera shoot can easily produce upwards of 100 terabytes of Original Camera Negative (OCN). Archiving this data is not simply a matter of copying it to a few external hard drives and placing them on a shelf. Hard disk drives (HDDs) have mechanical components that seize up over time, and their magnetic platters are susceptible to demagnetization (bit rot). Furthermore, solid-state drives (SSDs) will slowly leak their electrical charge if left unpowered for extended periods, resulting in catastrophic data loss.

To combat these vulnerabilities, the archiving workflow requires a "defense-in-depth" strategy. This involves migrating data away from active, spinning disks onto stable, offline media (like magnetic data tape), implementing mathematical verification algorithms to monitor data integrity over time, and relying on standardized, non-proprietary file formats. The ultimate challenge is designing an archive that is entirely self-describing—meaning the metadata and organizational logic are embedded directly alongside the media, ensuring future generations do not need a specialized database to decode the contents of the archive.

## 2. Archiving levels (working, vault, deep)

To manage massive amounts of data economically, post-production facilities implement a tiered storage architecture, dividing the archiving process into three distinct levels: Working (or Nearline), Vault (or Near-Archive), and Deep Archive.

The "Working Archive" represents data that is no longer actively being edited or graded every day, but might be needed at a moment's notice. For example, if a television series has finished its first season but the editors anticipate needing specific establishing shots for season two, those shots are moved from the hyper-expensive, high-speed flash SAN (Storage Area Network) to a slower, cheaper nearline NAS (Network Attached Storage). This tier prioritizes immediate accessibility over long-term stability.

The "Vault" tier represents the first true stage of long-term preservation. When a project is entirely finished, the assets are packaged and moved to this level. Typically, this involves writing the data to LTO (Linear Tape-Open) magnetic tape, but keeping those tapes inside a robotic tape library (a "tape silo"). If a client requests a localized master a year later, the robotic arm can automatically fetch the required tape, load it into a drive, and restore the data to the active servers within hours.

The "Deep Archive" is the final destination. This is for projects that have exhausted their commercial delivery cycles and are being preserved for historical record or future remastering (e.g., a 10-year anniversary 8K remaster). Deep archive data is written to multiple redundant LTO tapes, which are then physically removed from the robotic library and transported to geographically separated, climate-controlled, secure physical facilities. Deep archiving prioritizes maximum data security and longevity, accepting that retrieval may take days to physically transport and load the media.

## 3. Long-term archiving formats

A critical vulnerability in digital archiving is format obsolescence. If a project is archived as a proprietary Adobe Premiere Pro project file referencing raw RED .R3D media, opening that archive twenty years later relies on the assumption that Adobe software and RED decoding plugins will still exist and remain backward-compatible. To mitigate this risk, the archiving workflow prioritizes open, standardized, and uncompressed (or mathematically lossless) formats.

For image preservation, the gold standard is the Digital Source Master (DSM) or Master Animation (MA), typically rendered as an uncompressed 16-bit TIFF image sequence, or a DPX sequence. Image sequences are highly resilient; because every frame is a separate file, if a few bytes of data are corrupted on the storage medium, you only lose a single frame of video, rather than destroying the file header of a single monolithic video file (like a massive ProRes or MXF), which could render the entire film unplayable. More recently, SMPTE ST 2067 (the Interoperable Master Format, or IMF) has become a favored archival container. Because IMF uses component-based XML architecture referencing standardized JPEG 2000 image sequences, it provides a highly stable, non-proprietary framework for long-term preservation.

For audio preservation, standard uncompressed Broadcast WAV (BWF) files at 24-bit/48kHz or 96kHz are utilized. Crucially, the archive must include the final mixed stems (Dialogue, Music, Effects, Foley) entirely separated, rather than just the final composite mix. This ensures that future generations can create new immersive mixes (like a future format beyond Dolby Atmos) or replace the music score entirely without destroying the dialogue track.

## 4. Metadata and cataloging

An archive is entirely useless if the data cannot be found. Archiving 50 terabytes of ambiguously named files into a deep storage vault is the digital equivalent of dumping unmarked boxes into a warehouse. Therefore, metadata and meticulous cataloging are the most vital human elements of the archiving workflow.

Metadata must be embedded at multiple levels. At the file level, structural metadata (resolution, color space, frame rate) is typically baked into the headers of the DPX or IMF files. However, the archive also requires extensive descriptive metadata. This includes the project name, the director, the production company, copyright information, and a detailed manifest of the contents.

Before the data is written to long-term storage, a post-production coordinator generates an Archival Manifest. This is often a standardized XML or CSV document that lists every single file being archived, its directory path, its exact byte size, and its cryptographic checksum. Furthermore, facilities use dedicated MAM (Media Asset Management) software to index these archives. When an LTO tape is ejected and sent to a physical vault, the MAM retains a proxy-resolution viewing copy of the master and a searchable database of the tape's contents. If a producer calls five years later looking for a specific scene, the archivist can search the MAM, view the low-res proxy to confirm it is the correct shot, and the system will explicitly report: "This high-resolution file is located on LTO Tape Barcode #XYZ-123 in the deep storage vault."

## 5. Storage: LTO, cloud, and disks

The physical medium chosen for archiving dictates the cost, accessibility, and longevity of the data. Currently, the undisputed champion of deep archiving is LTO (Linear Tape-Open) magnetic tape. LTO offers a massive storage density (LTO-9 can hold 18TB of uncompressed data per cartridge), it is completely immune to the mechanical failures of spinning hard drives, and it has an estimated shelf life of up to 30 years if stored in a climate-controlled environment. Furthermore, because LTO tapes are "offline" (unpowered) when sitting on a shelf, they provide a physical "air gap" that makes them completely immune to ransomware or network-based cyberattacks.

To make LTO tapes behave more like standard hard drives, the industry utilizes LTFS (Linear Tape File System). LTFS partitions the tape, creating a small directory index at the beginning. This allows a user to mount the tape on a desktop computer and drag-and-drop files directly in the operating system's file explorer, relying on open-source standards rather than proprietary backup software.

Cloud storage (like Amazon S3 Glacier or Google Cloud Archive) has become an increasingly popular secondary archiving tier. Cloud providers utilize massive, geo-redundant server farms to ensure extreme data durability. The advantage of cloud archiving is that the burden of hardware maintenance and migration is entirely outsourced to the provider. The disadvantage is cost; while uploading data is cheap, "egress" fees (the cost to download the 50 terabytes back to your local facility) can be exorbitant. Finally, bare Hard Disk Drives (HDDs) are never recommended for deep archiving due to mechanical decay, but they are often used as temporary "shuttle drives" to physically transport the archive to the vault before the data is ingested to tape.

## 6. Integrity verification and migration

Data written to an archive is never assumed to be safe; it must be mathematically proven to be safe. In the archiving workflow, this proof is provided by hashing algorithms and checksums (such as MD5, SHA-256, or xxHash).

When the final master files are generated on the primary storage server, the archiving software calculates a unique alphanumeric string (the checksum) based on the exact binary composition of the file. If even a single byte of data changes, the resulting checksum will be radically different. When the files are copied to the LTO tape or the cloud, the software recalculates the checksums on the destination media and compares them to the originals. If the hashes match perfectly, the archive is verified as a bit-for-bit, flawless clone. These checksum manifests are saved alongside the media, allowing future archivists to verify that the files have not degraded while sitting in the vault.

However, data integrity relies on the survival of the reading hardware. LTO technology progresses in generations (LTO-6, LTO-7, LTO-8, etc.). The LTO consortium dictates that a new drive can only read tapes from one or two generations prior. Therefore, if a facility archives a film on LTO-5, and twenty years later the facility has upgraded entirely to LTO-9 drives, the old tapes will be physically unreadable. This necessitates the process of "Data Migration." Every 5 to 7 years, post-production facilities must undertake massive migration projects, pulling old tapes from the vault, verifying their checksums, and rewriting the data onto the newest generation of media, ensuring the archive outlives the hardware used to create it.

## 7. Master archive vs deliverables

A critical distinction in post-production is understanding the difference between the Delivery Masters and the Archival Master. Delivery masters are highly specific, heavily optimized files generated for immediate commercial distribution. For example, a ProRes 422 HQ file for broadcast television, an IMF package for Netflix, and an encrypted DCP for theatrical release are all deliverables. They are bound by the technical constraints, color space (like Rec.709 or P3), and compression limitations of their target platforms.

The Archival Master (or Digital Source Master - DSM), on the other hand, is the ultimate, uncompromised, highest-quality version of the project, completely untethered from immediate delivery constraints. It is typically rendered in the widest possible color gamut (like ACES AP0), at the maximum bit depth (16-bit float), and often in a textless format (without burned-in subtitles or title cards).

Why is this distinction necessary? Because delivery formats become obsolete rapidly, but the Archival Master must serve as the pristine source for all future unknown formats. If a film was finished in 2010, the deliverable was likely a 1080p SDR file. If that film was archived properly using a 4K, 16-bit ACES EXR Archival Master, it can be easily pulled from the vault today and remastered for a modern 4K HDR Dolby Vision release. If only the 1080p deliverable had been archived, generating a true HDR version would be mathematically impossible. The Archival Master protects the latent potential of the creative work.

## 8. Documentation and traceability

The physical and digital components of an archive are only half of the preservation equation. The other half is the documentation. Archiving a complex VFX-heavy feature film without thorough documentation is akin to handing someone a 10,000-piece puzzle without the picture on the box. In thirty years, the post-production supervisor, the colorist, and the editor will not be available to answer questions about how the project was structured. The archive must explain itself.

Comprehensive documentation includes a detailed "Post-Production Technical Bible." This document explicitly outlines the entire pipeline used to create the film: the specific versions of the editing software (e.g., Avid Media Composer 2023.8), the exact color science and LUTs utilized, the camera raw settings, the naming conventions, and the track layout of the audio stems (e.g., Track 1-2: Stereo Mix, Track 3-8: 5.1 M&E).

Furthermore, traceability must be maintained regarding the physical location of the media. This involves a rigorous chain-of-custody protocol. When an LTO tape leaves the facility, a barcode is scanned, and a digital ledger is updated indicating the exact courier used, the date of transit, and the specific geographic vault (and shelf number) where the tape was deposited. This traceability ensures that in the event of a studio audit, a corporate acquisition, or a remastering request, the legal owners of the copyright can instantly locate and retrieve their intellectual property.

## 9. Best practices and common pitfalls

The golden rule of archiving best practices is the "3-2-1 Backup Strategy." For critical, irreplaceable digital assets, facilities must maintain at least three copies of the data, stored on two different types of media (e.g., flash storage and LTO tape), with at least one copy stored securely off-site in a separate geographic location to protect against localized disasters like fires or floods.

A very common and dangerous pitfall is archiving proprietary project files (like a DaVinci Resolve .drp or Premiere Pro .prproj) without also rendering out high-quality, flattened media stems. Project files are merely XML databases linking to media; they are highly susceptible to software versioning issues. Best practice dictates that alongside the project file, the archivist must render a textless, flattened, high-resolution QuickTime or image sequence of the final timeline. This guarantees that even if the software used to edit the film ceases to exist, the finished linear video will survive.

Another frequent mistake is delaying the archiving process until months after the project has wrapped. Post-production facilities often face extreme pressure to clear up their expensive high-speed SAN storage for the next incoming project. Rushing the archive process leads to poorly cataloged data, skipped checksum verifications, and missing assets. Archiving must be treated as a formally budgeted and scheduled phase of the production pipeline, executed immediately upon the approval of the final deliverables, ensuring that the legacy of the film is secured with the exact same rigor that was applied to its creation.

In conclusion, the archiving workflow is the ultimate safeguard of cinematic and audiovisual history. It is a highly rigorous, defense-in-depth operation that transcends simple data duplication. By utilizing stable, long-term mediums like LTO tape via LTFS, employing mathematically flawless checksum verification, and insisting on open, non-proprietary image sequences, post-production facilities ensure that the creative endeavors of today remain accessible to the audiences and technicians of tomorrow. The archive is the final resting place of the production, but it must be built with the assumption of eventual resurrection. A properly cataloged, self-describing archive featuring an uncompromised, high-fidelity Digital Source Master allows a film to continually adapt to emerging display technologies, from SDR to HDR and beyond. Neglecting the archiving phase, or treating it as a hasty afterthought to clear server space, virtually guarantees that the intellectual property will succumb to digital rot or format obsolescence within a decade. To understand the precise technical rigor required immediately preceding the archive phase, explore our comprehensive guide to the 'Finishing workflow: from grade to masters'. To grasp the stringent parameters that these files must meet before they are considered final, review the 'Quality control and deliverable compliance' protocols. Finally, to see how massive amounts of data are managed collaboratively before they ever reach the vault, delve into our overview of 'The production pipeline: an overview'.

Sequential read speeds (MB/s) by storage tier. RAW video at 4K 24 fps requires ~1 Gbps for ProRes 422, ~5 Gbps for ARRIRAW.

## Related Articles

[Advanced

### Collaborative workflow: distributed teams and tools

Learn how collaborative processes and metadata generation ultimately feed the final archive.

Read article](https://postproduction.studio/en/workflows/collaborative-workflow)[Advanced

### Finishing workflow: from grade to masters

Learn how final deliverables are generated prior to long-term archiving.

Read article](https://postproduction.studio/en/workflows/finishing-workflow)[Fundamental

### The production pipeline: an overview

Understand how archiving sits at the absolute end of the production pipeline.

Read article](https://postproduction.studio/en/workflows/production-pipeline)

## Key terms in this article

* [Archiving](https://postproduction.studio/en/glossary/archiving)
* [LTO (Linear Tape-Open)](https://postproduction.studio/en/glossary/lto)
* [Checksum](https://postproduction.studio/en/glossary/checksum)
* [Pipeline](https://postproduction.studio/en/glossary/pipeline)
* [Deliverables](https://postproduction.studio/en/glossary/deliverables)
* [IMF (format)](https://postproduction.studio/en/glossary/imf-format)
* [XML (Extensible Markup Language)](https://postproduction.studio/en/glossary/xml-extensible-markup-language)
* [SAN](https://postproduction.studio/en/glossary/san)
* [ProRes](https://postproduction.studio/en/glossary/prores)
* [HDR (High Dynamic Range)](https://postproduction.studio/en/glossary/hdr-high-dynamic-range)
* [SDR (Standard Dynamic Range)](https://postproduction.studio/en/glossary/sdr-standard-dynamic-range)
* [NAS](https://postproduction.studio/en/glossary/nas)
* [Dynamic range (camera)](https://postproduction.studio/en/glossary/dynamic-range-camera)
* [ACES (Academy Color Encoding System)](https://postproduction.studio/en/glossary/aces-academy-color-encoding-system)