Common Mistakes When Scanning Family Photos

Instead: scan both sides of every print, even when the back looks blank.

The back of an old print is the single richest source of dating and context information you have access to: handwritten dates, names, locations, occasions, and faded developer timestamps. If you only scan the front, you throw all of that away: permanently, because you’re probably not going to feed every print through the scanner a second time.

Always scan both sides, even when the back looks blank. Many scanners have an auto-detect mode that only saves the back when text is visible: turn that off and force-save every back. Light pencil notes and faint timestamps are exactly the ones an auto-detector misses, and they’re the ones AI dating tools can still read. See our scanning best practices guide for the exact settings on the most common feed scanners.

Instead: scan prints at 600 DPI, and negatives or slides at 1200–2400 DPI.

300 DPI looks fine on screen at the print’s original size. Then your grandkid wants to enlarge a 4×6 of their grandmother to hang on a wall, or a restoration tool needs more pixels to repair a tear, or an AI upscaler tries to read the timestamp in the corner. At 300 DPI there’s not much to work with.

Scan at 600 DPI. Storage is cheap, hard drives keep getting bigger, and the prints themselves are not getting any younger. The professional archival standard is 600 DPI for prints and 1200–2400 DPI for negatives or slides. You only get one chance to do this with each photo: do it at the resolution you’ll wish you’d picked to last.

Instead: wipe the glass between batches and dust the prints before they touch it.

Every speck of dust on the glass becomes a black spot on every photo for the next hour. Every smear becomes a streak. You won’t notice while you’re scanning at speed: you’ll notice when you’re looking at the photos six months later and every print from the third stack has the same fingerprint in the upper left or a vertical line if you're using an auto feed scanner.

Wipe the glass with a microfiber cloth between batches; a can of compressed air works well too. If the prints themselves are dusty, give them the same treatment before they touch the glass.

Instead: pick one naming convention, like boxLabel_0001.jpg, and keep it for the whole project.

If batch one is named scan_0001.jpg, batch two is named IMG_0001.jpg, and batch three is named fastfoto_2024_05_09_0001.jpg, you have just made every downstream tool significantly more difficult.

Pick one naming convention and use it for the entire project. A practical one: <album-or-box-label>_<sequence>.jpg, e.g. Jul89Sept90_0001.jpg. The album label gives you a known time window per batch, the sequence number keeps front/back pairs together, and you can recognize at a glance which physical box a file came from.

If the box or album has no label, describe it: purpleAlbum_0001.jpg, flowerBox_0001.jpg. When a question comes up later, the filename walks you straight back to the original print.

Instead: scan one album or box at a time, in the order the prints were stored.

If you dump the entire shoebox into one stack and feed it through in random order, you have lost a huge amount of context that was free for the taking: photos that were stored next to each other in an album are usually from the same event, the same trip, or the same year. AI dating tools (and your future self) can use that adjacency as a strong signal.

Scan one album, sleeve, or grouping at a time. Keep the order the prints were stored in. Save each grouping to its own folder. The original physical organization is the closest thing to ground truth you’ll ever get.