Applications get rejected for the quality of the scan far more often than for the content of the document. A dark photograph with a shadow across the middle, taken at an angle, with the edge of a desk visible, is not an unreasonable thing to send — it is what a phone produces if you point it at a page and press the button. But it is also the kind of thing that comes back marked “illegible, please resubmit”, two weeks later.
The good news is that a phone camera genuinely is a competent document scanner. Modern sensors comfortably out-resolve what any form requires. The gap between a bad scan and a good one is almost entirely technique, and it takes about two minutes to learn.
Light is the whole game
Every other variable is a rounding error next to this one.
The dominant failure is self-shadow: you stand between a ceiling light and the document, lean over to frame it, and your own head and shoulders cast a gradient across the page. You will not notice on the phone screen, because your eyes adapt. The reviewer, looking at it cold, sees a document that is half grey.
Use a window. Daylight is bright, even, and does not flicker. Put the document on a table beside a window, not directly in a shaft of sunlight — direct sun creates blown-out highlights and a hard shadow from any wrinkle in the paper. Overcast daylight is close to ideal.
Put the light in front of you, not behind you. If you must use artificial light, position yourself so it comes over your shoulder from the side rather than from directly overhead.
Do not use the flash. A phone flash sits next to the lens, so it throws light straight at the page and reflects it straight back. On glossy paper this produces a white circle that destroys whatever it lands on. On matte paper it produces a hot centre and dark corners. There is no document that is improved by an on-camera flash.
Watch for mixed lighting. Half daylight and half fluorescent gives you a page that is blue on one side and yellow on the other, and automatic white balance cannot correct both. Turn the room light off and use the window alone.
Shoot square on
Hold the phone parallel to the page, directly above its centre, not leaning over from one side.
A page photographed at an angle comes out as a trapezium — wider at the bottom than the top, or skewed to one side. It is usually still readable, which is exactly why people send it. But it reads as careless, and on a form that will be processed automatically it can defeat the software entirely.
The practical difficulty is that standing directly over a document with a phone puts your body between the document and the light. Solve it by moving the document rather than yourself: put it near the edge of the table, stand at that edge, and let the light come from the far side.
If your phone has a grid or a level in the camera settings, turn it on. Lining the page edges up with the grid lines is much easier than judging by eye.
Fill the frame
Get close enough that the page occupies most of the picture, with just a small border.
Every pixel spent on your desk, your hand or the floor is a pixel not spent on the document, and cropping afterwards throws away resolution you already paid for in file size. A page filling the frame at 12 megapixels gives you far more usable detail than the same page occupying a third of the frame at the same resolution.
Beyond the technical point, a photograph of a page floating in the middle of a picture of somebody’s kitchen table is a poor first impression on an application.
Use document mode
Almost every current phone has one, and it is materially better than the standard camera.
On iPhone it lives in the Notes app — create a note, tap the camera icon, and choose Scan Documents. On Android it is usually in Google Drive under the plus button, or in the Files or Camera app depending on the manufacturer. Several phones now detect a document automatically and offer the mode without being asked.
What it does, and why it matters: it finds the page edges and crops to them, corrects the perspective so a trapezium becomes a rectangle, flattens the lighting so a shadow gradient disappears, increases contrast so text is properly black against properly white, and often converts to greyscale — which alone can cut the file size by half or more.
The output is a document that looks scanned rather than photographed, at a fraction of the size. If you take nothing else from this article, take this: use document mode. It closes most of the gap between an amateur scan and a professional one automatically.
Prepare the page itself
Small physical details, disproportionate effect.
Flatten it. A page fresh out of a folded envelope has creases that catch the light as dark lines. Press it flat under a book for ten minutes, or at minimum smooth it firmly by hand.
Use a contrasting background. White paper on a white table gives the edge detection nothing to work with, and document mode will fail to find the page. A dark table, or a sheet of dark card, makes the boundary obvious.
Remove staples and paper clips before photographing. They cast shadows and lift the page.
Do one page at a time. Two pages side by side means each gets half the resolution, and most processing expects one page per image.
Check before you leave the desk
The most common avoidable mistake is discovering a bad scan after the document has been returned to its folder, its owner, or its filing cabinet.
Before you move on, zoom in on the smallest text on the page — the footnote, the account number, the reference code — at full magnification. If you cannot read it comfortably on the phone, nobody will read it on a screen either. Retake it there and then, while the document is still in front of you.
Also check the corners. Perspective correction sometimes clips a corner, and the corner is where reference numbers and stamps tend to live.
Turning the photographs into a document
Once you have your images, three steps produce a submittable PDF.
Combine them in the right order. Convert the images to a PDF with one image per
page. Get the order right before you convert — file browsers sort alphabetically, which puts
IMG_10 immediately after IMG_1, and fixing the sequence afterwards is more work than setting it
first.
Choose the page size deliberately. Fit to image keeps each page exactly the shape of its photograph, which is right for anything read on screen. A4 or Letter places each image on a standard sheet, which is what you want if the document will be printed, and what some portals actually validate before accepting a submission.
Fix any sideways pages. Phones record orientation as a tag rather than by rotating the pixels, and that tag is frequently ignored when the image goes into a PDF — which is why a photograph that looked correct in your gallery appears on its side in the document. Rotating the PDF writes the correction into the file permanently, so every reader shows it correctly.
Compress last, if you must. Photographs keep their full resolution in the PDF, so a five-page scan can easily exceed 20 MB. If there is an upload limit, compress the finished document once, at the end. Starting at 150 DPI with greyscale enabled will usually bring a phone scan under 2 MB while remaining perfectly readable.
When to use a real scanner
A phone is not always the right tool.
For anything over about twenty pages, a document feeder is worth the trip — twenty individual photographs is fifteen minutes of tedium and one of them will be wrong. For material that must be archival, or where fine detail genuinely matters, a flatbed at 600 DPI captures things a phone cannot. For fragile or valuable originals, a flatbed does not require handling and repositioning.
For everything else — the signed form, the utility bill, the certificate, the receipt, the ID page — a phone in document mode, with decent light, is entirely sufficient, and it is available right now.
Ready to assemble your scans? JPG to PDF combines them into one document, in your browser, with nothing uploaded.