Skip to content
Try Raven →
All posts
PlaybookBy the Raven team6 min read

Digitizing and Exploring Old Travel Albums

A practical guide to scanning old prints and slides properly, then using AI to help place the trips nobody wrote a caption for.

Short answer

To digitise old travel photos, scan prints on a flatbed at 600 dpi or more and scan slides or negatives with a film scanner at 2,400 dpi or higher. Save masters as TIFF, capture any writing on the back, then upload a copy to an AI tool for a location estimate.

Abstract cluster of faded passport-stamp shapes and a thin dashed flight-path line curving across a soft gradient, no text.

Somewhere in a cupboard, a box or a parent's loft, there is probably an album from a trip that predates digital cameras entirely — thick paper pages, little plastic sleeves, perhaps a few slides that need holding up to a lamp to see properly. These photographs carry a specific kind of mystery that camera-roll photos do not: no timestamp buried in the file, no GPS tag, sometimes not even a date written on the back. If nobody who remembers the trip is still around to ask, working out where a print was taken becomes real detective work.

This is a more hands-on version of the curiosity that drives casual AI geolocation, except that the first step is not uploading a photo. It is getting a decades-old print into a form a computer can look at properly.

Why are printed photos harder to place than digital ones?

A digital file usually carries a capture date even when its location field has been stripped. A print carries nothing machine-readable at all, so everything has to come from the paper itself, whatever is written on the back, and the content of the image.

A digital photograph arrives bundled with metadata, and even when it has been removed for privacy a rough date is often recoverable from file properties or backup timestamps — the piece on why phones add GPS data to photos covers what those fields hold and where to read them. A print has none of that. What you get instead is whatever handwriting made it onto the back, the manufacturer markings and lab stamps in the border, and the visual content of the photograph itself. That last part is where digitising pays off, because a phone snapshot of a print is a poor substitute for a proper scan the moment you want to zoom into a background detail.

How do you digitise old travel photos properly?

Clean the surface, scan prints on a flatbed with the lid closed, and use a film scanner or transparency adapter for slides and negatives. Scan the back whenever anything is written on it, and keep a lossless master before making smaller copies to share.

A flatbed scanner remains the most reliable option for prints. Slides and negatives need a dedicated film scanner or a transparency adapter, because a flatbed lamp alone produces flat, low-contrast results from film. Whichever device you use, the order of operations matters more than the price of the hardware.

  • Clean the print or slide first with a soft, dry microfibre cloth. Dust and fingerprints get scanned along with the image and are tedious to remove afterwards.
  • Close the lid and control the light. Scanning with the lid open under a ceiling light adds glare and a colour cast that no amount of correction fully removes.
  • Scan the back too whenever there is handwriting, a date stamp or a lab processing sticker. That context is the first thing lost once photos leave their album pages.
  • Keep a lossless master. Archive as TIFF and export JPEGs from it for sharing, rather than re-saving a JPEG repeatedly and compounding the compression each time.
  • Name the file descriptively, even if the guess is rough. Something like europe-trip-maybe-1974 is far more useful in five years than a scanner's default numbering.

What resolution do you actually need?

300 dpi is enough only to view a print at its original size. 600 dpi is the sensible floor if you intend to zoom into signage or architecture. Film carries much more detail, so scan 35 mm slides and negatives at 2,400 dpi or higher.

Resolution is the decision people most often get wrong, because it feels like a storage question and is really a detail question. Scanning is measured in dots per inch, and the arithmetic is unforgiving: a 35 mm frame measures 36 by 24 mm, so scanning it at 3,000 dpi yields an image of roughly 4,250 by 2,830 pixels. The same frame at 600 dpi gives about 850 by 570 pixels — fine on a phone screen, useless the moment you want to read a shop fascia in the background. Since a scanning session is tedious enough that nobody wants to repeat it, err upwards.

Film stock is also a dating clue in its own right. Slide mounts usually carry a processing month and year, and the stock itself narrows the range: Kodachrome was discontinued in 2009 after 74 years, so a Kodachrome slide has a hard upper bound and a very recognisable colour rendering. Between the mount, the stock and the clothing in the frame, a decade is usually recoverable before any software is involved.

What can AI add once a print is scanned?

Once digitised, a print is an ordinary image file and can be analysed like any other. Raven reads the architecture, vegetation, road markings and signage in the scan and estimates a region — an entertainment-only guess that is often enough to prompt a relative's memory.

Upload the scan at withraven.net and Google's Gemini model reasons about what is visible in the frame rather than looking anything up. It knows nothing about the family trip, obviously, but a guess such as "likely northern Italy", paired with a decade suggested by the film stock and the clothing, is often enough to jog a relative into filling in the rest. The image is held in memory for that single request and never written to disk or a database. The web version gives one free guess per account and then points to the free Geospy AI app for iPhone, which is easier to use when the scanner is in another room. If you have not run the flow before, the step-by-step walkthrough explains how to read the confidence indicator.

Making it a project worth finishing

This works best as a slow, ongoing project rather than a weekend sprint. A box of a few hundred prints is a great deal to scan in one sitting, and the interesting part is usually the handful of unlabelled photographs that need real detective work anyway. Digitising also makes the results shareable: a scanned, dated and roughly placed photograph can go into a family group chat in a way a fragile paper original never could.

Two habits are worth carrying through the whole exercise. Keep the location precision you publish separate from the precision you store, exactly as the geotagging guide recommends for new work. And remember that an inherited album is other people's history as well as yours — the piece on AI and photo privacy sets out where the line sits. Digitising the past turns out to be the easy half; placing it is the part that stays interesting.

Scan one unlabelled print and see what Raven makes of the background.

Upload a photo →

Frequently asked questions

What scanning resolution should I use for old prints?
600 dpi is a sensible floor for prints you may want to zoom into later, and 300 dpi is enough only for viewing at original size. Film needs far more: 2,400 dpi is a practical minimum for a 35 mm slide, and 3,000 to 4,000 dpi is better.
Can a phone photo of a print replace a scan?
For a quick share, yes. For analysis, no. A handheld copy adds glare, keystone distortion and camera shake, and it softens exactly the small background details — lettering, kerbs, roof tiles — that help place the scene.
How do I date a print with nothing written on the back?
Three clues usually converge: the film stock or processing lab stamp printed in the border, the paper format itself, and the clothing and vehicles in the frame. Each gives a range, and the overlap is often narrow enough to name a decade.
Is it safe to upload inherited family photos?
Raven processes an uploaded image in memory and never stores it, but the courtesy question still applies. If a photograph shows other people, treat it as theirs as well as yours and ask before you post the result anywhere public.

Sources

  1. Dots per inchWikipediaExplains the scanning measure behind the 300, 600 and 2,400 dpi figures used here.
  2. TIFFWikipediaThe lossless container, first published in 1986, normally used for archival scan masters.
  3. KodachromeWikipediaThe colour reversal film discontinued in 2009, whose distinctive rendering helps date a slide.

Reminder

Raven is built for entertainment and curiosity. Its guesses are AI estimates that can be wrong, and it must never be used to track or identify real people. Uploaded photos are processed in memory and immediately discarded — never stored.

Get Geospy AI for iPhoneDownload free