Evidence
Original capture or received file
The untouched camera, scanner, phone or third-party file. It may contain pixels, EXIF data, timestamps, colour information and source evidence that later versions no longer retain.
Long-term digital preservation is the continuing management of collection photographs so that they remain intact, authentic, understandable, discoverable and usable despite storage failure, software change, account loss and the passage of time. It is not the same as keeping old files, buying more cloud storage or running an occasional backup.
For a serious collector, photographs can become part of the collection record itself. They may document identity, condition, provenance, restoration, completeness, serial numbers, signatures, insurance evidence and the state of an object before sale, loss or damage. Preservation therefore protects more than pixels: it protects the continuing relationship between the image, the owned item and the evidence that gives the image meaning.
Preservation target
A usable photographic archive is a compound record. It includes the image data, the source file, any preservation master, useful derivatives, metadata, item links and enough preservation history to explain what has happened to the files. Treating the image folder as the whole archive leaves the most fragile layer—the meaning—outside the preservation plan.
Evidence
The untouched camera, scanner, phone or third-party file. It may contain pixels, EXIF data, timestamps, colour information and source evidence that later versions no longer retain.
Longevity
A stable, high-quality version retained for future rendering, migration or reproduction. It may be the native RAW, a validated DNG, a simple TIFF or the best available original JPEG.
Use
Crops, corrected copies, thumbnails, marketplace files and web images should be generated from protected sources rather than becoming the only surviving versions.
Meaning
Identifiers, captions, view types, rights, provenance and links to the specific owned item are part of the preservation target. Pixels without context can survive while the record still fails.
Boundary
Backup mainly protects against loss and permits recovery to an earlier state. Preservation additionally requires an inventory, checksums, metadata, exportable relationships, format awareness, permissions, storage migration, documented procedures and tested restoration. Three copies of the same corrupted or unintelligible archive are still three failed copies.
Risk diagnosis
Accidental deletion, failed media, account closure and synchronised deletion can remove every convenient copy at once.
Corruption, interrupted transfers, ransomware or software faults may alter a file without producing an obvious visual warning.
A file may remain intact while its format, catalogue software, storage interface or proprietary export becomes difficult to use.
IMG_4827.JPG may be clear today and meaningless to an executor, buyer or future researcher twenty years later.
Photographs can reveal locations, serial numbers, alarms, personal documents and high-value holdings. Public access and preservation access should not be treated as identical.
An archive can be technically sound yet unusable because nobody knows where it is, which copy is authoritative, how to restore it or who can access the credentials.
Formats
Preservation format decisions should consider openness, documentation, adoption, metadata support, compression, colour management, software support and migration options. There is no rule that every collector must convert every image to TIFF. Retain what you received before normalising it, and create additional formats only when they serve a defined preservation or access purpose.
Native source
Retains extensive sensor data and editing latitude, but may depend on manufacturer-specific decoding. Preserve the native RAW for significant work even when a rendered master is also created.
Normalised RAW
Broadly supported and publicly documented, but conversion should be validated. For important files, retaining both the proprietary RAW and the DNG may be justified.
Rendered master
Mature, high-quality and widely supported. Prefer simple, well-documented TIFFs without unnecessary layers, exotic compression or proprietary features.
Interoperable original or access copy
Lossy but extremely well supported. An original JPEG should be preserved as an original; converting it to TIFF does not restore information already lost through compression.
Graphics and annotations
Useful for screenshots, diagrams, transparent backgrounds and annotated details. It is not normally a replacement for RAW capture of ordinary photographic subjects.
Phone capture
Efficient and common, but less universally supported. Retain the native file and consider creating a validated, broadly renderable companion copy.
Fixity
A checksum is a value calculated from a file's contents. If the contents change, the checksum will almost certainly change. SHA-256 is a practical, widely supported choice for collector archives. Use checksums when files enter the archive, after copying, before and after storage migration, after restoration and during scheduled integrity audits.
| A checksum can show | A checksum cannot show |
|---|---|
| Whether a copied file is byte-for-byte identical | Whether the photograph was truthful when created |
| Whether silent corruption or alteration occurred | Who created the image or owns its copyright |
| Which backup contains the known-good copy | How to repair damage without another verified copy |
Copy strategy
A strong baseline is three independent copies, on at least two storage systems, with one copy geographically separate. Add one offline, immutable or strongly versioned copy and aim for zero unverified backup errors. The numbers are useful only when the copies do not share the same failure modes.
Migration
Move the same files from an ageing drive, NAS, optical disc or cloud provider to a current storage system.
The file contents should remain unchanged and be confirmed through checksum comparison.
Transform content from one format to another, such as HEIC to TIFF or a proprietary catalogue export to CSV.
The checksum will change, so significant properties such as dimensions, colour profile, bit depth, orientation, transparency and metadata must be validated.
Meaning and context
The image archive and the collection database form a compound record. A durable export should include items, variants, owned copies, image records, image-to-item relationships, captions, subjects or view types, original filenames, rights, timestamps, checksums, status flags and controlled vocabulary definitions. A pile of images is not a complete export of a digital collection.
/archive/
/images-original/
/images-master/
/images-access/
/metadata/
items.csv
images.csv
image-item-links.csv
controlled-vocabularies.csv
/manifests/
sha256-manifest.txt
/documentation/
README.txt
preservation-policy.pdf
data-dictionary.csv
export-notes.txtWorkflow
Copy files from the camera, scanner, phone, seller or auction source without editing the only copy.
Confirm expected file count, openability, completeness and malware risk where relevant.
Place received or captured source files in a protected area and treat them as read-only evidence.
Connect each significant image to the owned item, image role and photographic set.
Record subject, view, date, creator, source, rights and purpose while the context is fresh.
Create a fixity baseline, commonly using SHA-256, before files are copied onward.
Produce preservation-quality renders where justified and disposable access versions for routine use.
Copy to independent local and off-site storage, then verify the transfer rather than assuming it succeeded.
Mark the ingest as complete so files do not remain indefinitely in an unresolved incoming folder.
Collector judgement
Highest priority
Unique photographs, condition evidence, authenticity details, provenance records, pre-restoration views, lost or sold objects, rare variants and legally or financially significant images.
Medium priority
Routine collection views, repeated but useful detail sets, display and exhibition history, event documentation and research working sets.
Lower priority
Regenerable thumbnails, duplicate exports and temporary marketplace crops—after confirming that they carry no unique context or publication history.
Maturity
Level 1
Two copies exist, one is separate from the working computer, filenames are understandable, essential images are linked to items and a basic catalogue export exists.
Level 2
At least three independent copies exist, one is geographically separate, originals are separated from derivatives, backups are automated and restoration is tested.
Level 3
Checksums, manifests, periodic fixity checks, format identification and verified storage migrations provide evidence that files have not silently changed.
Level 4
Format risk is monitored, metadata and relationships are openly exportable, procedures are documented, immutable protection exists and succession arrangements are workable.
Maintenance
Security and succession
Collector photographs may expose home interiors, storage locations, serial numbers, personal papers and high-value holdings. Use least-privilege access, multi-factor authentication, separate administrator credentials, encrypted transport and storage, and reduced-metadata public derivatives. Keep fuller private evidence only where justified.
Common failure patterns
Myth
It is in the cloud, so it is preserved
Reality
Cloud accounts can be deleted, altered, locked, discontinued or left inaccessible after death.
Myth
RAID means I have a backup
Reality
RAID improves availability after some disk failures; it does not create an independent historical recovery copy.
Myth
The files open, so they are healthy
Reality
Visible opening is not a substitute for checksum comparison, validation and tested restoration.
Myth
My collection app stores everything
Reality
The application may omit originals, embedded metadata, relationships, deletion history or complete exportability.
Myth
RAW is the original, so I need nothing else
Reality
RAW may require specialist interpretation. A rendered master, metadata and an open catalogue export may also be essential.
Myth
I will organise it later
Reality
Context decays faster than files. Identity, source and purpose should be recorded during ingestion.
Minimum standard
Understand how to separate evidential source files from later working and access versions.
Return to the full image-management chapter and its practical topic sequence.
Continue into the wider storage decisions that protect both physical objects and their records.
Build copies that do not all share the same device, location, account or failure mode.
Preserve the descriptive and evidential context that makes an image intelligible.
Keep image files connected to the correct item, owned copy, condition state and documents.
Separate preservation access from public sharing and control exposure of sensitive collector information.