Scanning documents at home: resolution, PDF and the good enough rule

Most households now have a flatbed scanner, usually bolted to the top of a printer. Most use it twice a year and every time wonder why the scan looked blocky on the screen or weighed thirty megabytes for no reason. Here is a practical guide to settings that make scans useful.

How a flatbed scanner works, in one paragraph

A lamp slides under the glass, a lens focuses the reflected light onto a line of sensors, and the machine stitches the lines together as the lamp moves. Older flatbeds use a CCD and a mirror; most modern home scanners use a contact image sensor that sits right against the glass. Either way, the output is pixels. More on the underlying hardware in the Wikipedia overview of the image scanner.

Pick the right resolution

Scanner resolution is in dots per inch (DPI). The DPI number is the detail per inch of the original, not per inch of your screen. For most household originals:

Doubling the DPI roughly quadruples the file. Pick the lowest value that still looks right.

150 dpi is plenty for a receipt you just need to keep for an insurance claim.
300 dpi is the standard for anything you want to read clearly or print again at the same size. Letters, meeting notes, forms.
600 dpi is the right choice for photos you intend to keep, old documents with faint ink, and anything you might want to enlarge later.
1200 dpi and above is for small originals (a stamp, a small slide) where you want to recover detail you cannot see.

Doubling the DPI quadruples the file. Start at 300 and only go up if the result on screen is clearly fuzzy.

PDF or JPEG?

A quick rule:

Documents (letters, forms, invoices, anything with text) go to PDF. The file compresses well, multiple pages stay in one place, and PDF is the format everyone else can open without thinking.
Photos go to JPEG for sharing (small, universal) or to PNG/TIFF if you want to edit them later. JPEG is lossy; keep the master if you plan to touch it up.

If you are scanning something to keep for a long time, such as a solicitor's letter or a certificate, PDF/A is the archival flavour of PDF, designed to open the same way decades from now. Many scanner drivers have a PDF/A option tucked in the advanced settings.

Black and white, greyscale or colour?

Black and white (one bit per pixel) is for text-only documents where you want the smallest file. Modern scanners handle it well; older ones can produce speckled output.
Greyscale (256 shades) is almost always a better choice for everyday text, including anything printed in black ink that includes a stamp or signature. The file is still small.
Colour is for photos, logos, and anything where the colour matters for the content, including those green and yellow "important" stamps on official letters.

Flatten the lid, square the page

Three small things save a lot of scrolling:

Push the lid down firmly before scanning a book or a thick card. Light leaks around the edges ruin a scan.
Line the paper up with the corner marks on the glass. Scans that come out at 3° off vertical are a daily annoyance.
Use the auto-crop option if your driver has one. It saves time and avoids a black border down one side.

When to use a phone instead

For a single-page receipt that you just need to email, the camera in your phone is faster than firing up the scanner. The built-in scanner in iOS Files and the Google Drive app both deskew, convert to greyscale, and save as PDF. They are not the right tool for a stack of forms, but for one page they are quicker and the result is usually good enough.

OCR: making scans searchable

Optical character recognition turns the image of a letter into selectable text. If your scanner software offers "Searchable PDF", that is OCR; the file still looks like a photo of the page but you can copy sentences from it or search inside it. Enable it for anything you might need to find later. The quality depends on how clean the scan is, which brings us back to resolution and squareness.

Keeping your scans somewhere

Scans are small files that quickly add up. A simple year-and-month folder structure on your phone or a laptop will beat any clever system you will not maintain. If you use a cloud drive, remember that scans of utility bills, prescriptions and certificates sit next to your family photos: pick a drive you trust, and consider a password on the folder.

If you are scanning a lot of family photos, our paper and duplex guide has notes on how a good scan goes back out to print well.

Common faults that make scans look bad

A quick check-list for when a scan comes out worse than the original:

Dirty glass. A fine mist of kitchen grease or a smear of fingerprints will show up as grey streaks on every scan. Wipe the glass with a lightly damp microfibre cloth and dry it.
A speck of dust can show up as a hair or a dot in the middle of the page. If you keep seeing the same mark in the same place, open the lid and clean it off.
Colour cast on old photos. Many scanner drivers have a "restore faded colours" option. It is often too aggressive; try it, and if the result is cartoonish, turn it off and bring the photo back manually later.
Moiré patterns when scanning newspapers or magazines. The pattern is the printer's halftone dots fighting the scanner's sensor grid. Enable the driver's descreen option or scan at a different resolution.

A note on sensitive paperwork

If you are scanning passports, visas, benefit letters or payslips to send to a solicitor, a landlord or an estate agent, use a password-protected PDF rather than a raw JPEG. Most scanner drivers can set a password on the saved file. It will not stop a determined actor, but it will stop an email reaching the wrong inbox from being trivially useful.

If you are scanning for a visa application or similar process, check what the receiving body wants: some will only accept colour PDFs at a specific resolution, and a scan that is too high can be rejected just as a scan that is too low can.