Short answer: use a scanner app rather than the camera, light the page from the side rather than overhead, and export as PDF in greyscale. That combination produces something that reads as a scanned document rather than a photograph of paper.
Why a photo is not a scan
A camera photo of a document has perspective distortion, uneven lighting, a visible background and a colour cast. A scan has square edges, flat white background and high contrast text. Scanner apps do four things automatically to bridge that gap:
- Edge detection — finds the page boundary and crops to it
- Perspective correction — squares up the page even if you shot it at an angle
- Contrast normalisation — pushes the background to white and text to black
- Multi-page assembly — collects pages into one PDF
The built-in options
Neither platform needs an install for basic scanning:
iPhone: Notes → new note → camera icon → Scan Documents. Also available in Files. Auto-captures when it detects a page.
Android: Google Drive → New → Scan. On Pixel and many Samsung devices, the Camera app also offers a document mode.
These are fine for a one-off. They fall short on batch work, they bury the output in an app you may not want it in, and their OCR export is limited.
Getting the lighting right
This matters more than the camera. Three rules:
- Light from the side, not above. Overhead light means your phone casts a shadow across the page.
- Avoid mixed light sources. Daylight through a window plus a warm desk lamp produces a colour gradient across the page that no filter fixes cleanly.
- Kill the glare. Glossy paper and laminated cards reflect. Tilt the page slightly rather than the phone, so the reflection moves off-axis.
If you must shoot under a single overhead light, hold the phone off to one side and let perspective correction square the result.
Colour, greyscale or black and white
| Mode | Use for | Typical size (A4 text page) |
|---|---|---|
| Black & white | Plain text, forms, receipts | ~50–150 KB |
| Greyscale | Text with photos or diagrams | ~200–500 KB |
| Colour | Anything where colour carries meaning | ~1–3 MB |
Default to black and white for text. It produces the smallest file, the highest contrast, and the best OCR accuracy. Switch to colour only when colour is actually information — a highlighted contract clause, a colour-coded chart, an ID card.
Resolution: how much is enough
- 150 DPI — screen reading only, do not OCR
- 200 DPI — the practical minimum for reliable OCR
- 300 DPI — the standard for documents you may print or archive
- 600 DPI — only for small print or images you intend to enlarge
Higher is not better. A 600 DPI colour scan of a text page is a 20 MB file that no email server will thank you for and that OCR reads no more accurately than the 300 DPI version.
Making it searchable with OCR
There are two different things people mean by OCR:
- On-screen text selection — iOS Live Text and Google Lens let you select and copy text out of an image. Nothing is saved.
- Searchable PDF — the scan image stays visible, with an invisible text layer behind it. The file is searchable, selectable and indexable.
For anything you are archiving, you want the second. A document scanner app exports searchable PDFs directly; otherwise run the finished PDF through an OCR tool.
Practical checklist
Before you send it:
- Pages in the right order, none missing
- Every page fully in frame — check the corners
- Text legible when zoomed to 100%
- File under 10 MB if emailing, under 5 MB for most web upload forms
- Filename that means something —
invoice-2026-09-acme.pdf, notscan_0001.pdf
When you need a real scanner
A phone will not match a flatbed for: bound books that will not lie flat, very large formats, photographs where colour fidelity matters, or bulk jobs of more than about 50 pages. For everything else — receipts, forms, contracts, ID documents, homework — a phone with good lighting matches a consumer flatbed closely enough that the recipient will not be able to tell.
