Photo digitization formats and disc handling rules for a lasting family archive

What to take away

Choosing image and document formats

The University of Michigan Library's digital archiving guidance on text and images explains the reasoning archivists use. It admits software keeps changing, so no format in use today is guaranteed accessible in ten years.

Its examples are usable directly. Where only the content of a document matters, plain text or csv are simple and stable. Where the layout has to survive, pdf is an acceptable standard because it is widely used and can be rendered by many programs and browsers.

For images the same three tests apply. TIFF is the usual master format: an open specification, stored uncompressed or with lossless compression. JPEG is the usual derivative: compressed, lossy, and readable by nearly every program. Decide once, for the whole collection.

This article quotes no prices. The linked guide covers what a digitization project costs.

Media, and the disc rules that matter

If discs are part of your plan, the handling makes a real difference. The NIST guide for librarians and archivists on the care and handling of CDs and DVDs contains a quick reference list that is more specific than most advice on the subject.

Its do-not list covers routine household actions, starting with these: do not use adhesive labels, write or mark in the data area, or scratch the label side. Do not use a pen, pencil, or fine-tip marker on a disc, or markers containing solvents.

Do not try to peel off or reposition a label, and do not clean by wiping around the disc rather than across it. Do not store discs horizontally for years.

Its storage recommendation is a range, not a point. For CDs and DVDs stored together it gives less than 20C but greater than 4C, at 20 to 50 percent relative humidity.

It states that 18C at 40 percent is suitable for long-term storage, with lower figures for extended-term storage. For archiving recordable discs, it recommends discs with a gold metal reflective layer.

The label-side warning surprises people. On a CD the recording layer sits close to the label surface, so a scratch there is more serious than a scratch on the side the laser reads.

The file that makes the rest usable

Every folder of scans should hold a plain text file that nothing needs to open. Say what it holds, where the originals came from, when the scans were made, at what settings, and what the naming pattern means.

Plain text survives everything, so keep it in the folder, not a database, because the folder is what gets copied.

This is the cheapest insurance in the whole exercise. A collection with a readme can be understood by someone who was not there. Without one it is a directory of numbers.

A workable household plan

Two different media types, one of them offsite, both replaced on a schedule rather than at failure. Masters in an open format, derivatives in a common one. A plain text file in the same folder saying what everything is and when it was made.

A typical rotation is every three to five years, or sooner if a drive reports errors.

Our guide to what digitizing costs after the scanner covers the running side, and the notes on digitizing mistakes cover the failures this plan is designed to survive.

Common questions