How to get cleaner scanned PDFs from a phone

Light is the biggest factor

Even, diffuse light does more for a scan than any camera setting.

• Watch your own shadow: The most common mistake is standing directly under a ceiling lamp, which throws the shadow of your phone and hand straight onto the page. Move so the main light comes from the side or from in front, and the shadow falls away from the paper.
• Daylight, not direct sun: Daylight near a window works well, but avoid direct sunlight, which creates harsh contrast and bright spots.
• Flash on glossy paper: On glossy paper, like photographs, brochures and some receipts, flash produces glare that can wipe out whole words, so switch it off and add ambient light instead. On matte paper in dim rooms, flash can help.
• Don't mix light colors: Avoid mixing warm lamps with cool daylight when you can, because uneven color makes backgrounds blotchy and text harder to separate from paper.

Background, flatness and framing

• A contrasting surface: Put the page on a plain surface that contrasts with it: dark wood, a desk mat or a sheet of dark card behind white paper. A contrasting background makes the edges of the page obvious, both to you when framing and to anyone cropping later.
• Flatten creases and curls: Flatten creases and folds, and weigh down pages that curl, keeping fingers and weights outside the text.
• Bound books are hardest, because pages curve toward the spine and text bends with them. Pressing the book open flat helps.
• Fill the frame: Fill most of the frame with the page while keeping all four corners in view, and photograph one page per image.
• Clear the desk: Remove paper clips, sticky notes and anything else that isn't part of the document, because the camera records everything on the desk.

Hold the phone parallel to the page

• What tilting does: When the phone is tilted, the page looks like a trapezoid in the photo. Lines of text lean toward each other, letters on one edge look smaller than on the other, and the page looks crooked even if the paper is straight on the desk. This is called keystone or perspective distortion, and it makes text harder to read and to recognize.
• Shoot from directly above: Hold the phone directly above the center of the page with the screen parallel to the paper. Many camera apps can show grid lines that make this easier.
• Small rotations are fixable: A small rotation, where the page is level but turned a few degrees, can be corrected afterwards with Deskew PDF.
• Perspective distortion is not: Perspective distortion can't be corrected by VelloDoc's tools, so it's worth retaking the photo straight on.

Sharpness: focus and stillness

A slightly blurred photo can look acceptable on a phone screen and still defeat text recognition.

• Tap to focus: Tap the page on screen to focus before taking the picture, and wait a moment for the camera to settle.
• Steady the phone: Brace your elbows on the table or hold the phone with both hands to reduce shake.
• Light helps sharpness too: Good light helps here too: in bright conditions the camera uses a faster shutter, which freezes small movements, while in dim light it slows down and picks up every tremor.
• Check before you move on: After capturing a page, zoom in on the smallest text before moving on. Retaking a blurry photo takes seconds while the document is still in front of you, but discovering the problem later often means finding the paper again.

Resolution and file size

Modern phone cameras capture far more detail than a page needs.

• What a phone already gives you: A typical 12-megapixel photo is about 4,000 by 3,000 pixels, which across an A4 page works out to roughly 340 to 360 pixels per inch, more than enough for sharp reading and accurate OCR. As a rule of thumb, around 300 pixels per inch is plenty for ordinary printed text.
• The flip side is size: each photo can take several megabytes, so a ten-page scan quickly becomes too large to email.
• What Scan to PDF does: Scan to PDF shrinks very large images to 6,000 pixels on their longest side. Pages are sized at 150 pixels per inch, so they look big in a viewer but print correctly with fit to page.
• Compress at the end: Reduce the finished file with Compress PDF when size matters.

Color, grayscale or black and white

Capture in color and decide later, because color information can't be recovered once discarded.

• Grayscale for ordinary paperwork: For ordinary paperwork, converting the finished PDF to grayscale with Grayscale PDF usually cuts file size noticeably without losing anything that matters.
• Keep color when it carries meaning: official stamps and seals, signatures in blue ink that show a document is an original, highlighted passages, colored charts and photographs.
• Pure black and white, where every pixel is either black or white, makes the smallest files. But it can wipe out faint pencil marks, light stamps and gray shading.
• Why grayscale is safer: That's why grayscale is the safer choice for documents you may need to rely on later.

Cleaning up after capture

Once the photos are taken, a short sequence turns them into a clean document.

• Combine in the right order: First combine them in the right order with Scan to PDF, dragging the file list until the pages read correctly.
• Remove blank backs: If you photographed blank backs or separator sheets, remove them with Remove Blank Pages.
• Straighten tilted pages: Then straighten tilted pages with Deskew PDF, before OCR, so the text is read from level lines. Deskewing turns the page rather than redrawing it, so the scan keeps its detail.
• Trim borders: Trim dark borders and desk edges with Crop PDF, bearing in mind that cropping hides instead of deletes the trimmed area.
• Add a text layer: Finally, add a searchable text layer with OCR PDF, and compress the result last.
• Why the order matters: Following this order avoids repeating work and keeps quality where it matters.

Making the scan useful later

A scan is only valuable if you can find it and trust it.

• Confirm recognition worked: After OCR, search the PDF for a few words and numbers you can see to confirm recognition worked, paying special attention to names, dates and amounts.
• Name and title the file: Give the file a descriptive name, and set a clear title with View / Edit Metadata so it's easy to identify in a document system.
• Archive long-term records: For records you must keep for years, convert the finished file with PDF to PDF/A.
• What a photo can carry: Also remember what a photo can carry: phone pictures may include location data, and standard JPEG photos are embedded in the PDF with their original data. The PDF privacy guide explains how to share scans safely, and what makes a PDF searchable covers OCR in depth.
• iPhone HEIC photos: Photos taken on an iPhone are often HEIC files. The guide to turning HEIC photos into one PDF covers that case.

Questions

Tools for this topic