Skip to main content

Photographing Your Pages

If you're photographing pages of a register, ledger, or daybook instead of scanning, follow these guidelines to get clear, usable results.

Before you start​

  • One photo per page — or per open spread if the original is a bound volume and the spread is the natural unit.
  • Keep pages in order — number them as you go, or keep them in a neat stack. You can reorder them in the upload wizard, but starting in order saves time.
  • Phone cameras are fine — a modern smartphone takes perfectly usable photos. You don't need a scanner.
  • Don't crop tightly — include the margins. Page numbers, reference codes, and annotations often live in the margins; we use them for verification.

Getting a clear photo​

  1. Straight on. Hold the camera perpendicular to the page — not at an angle. If you're photographing a bound book, open it flat and shoot from directly above.
  2. Good light. Natural daylight is best. Avoid shadows crossing the page. If you're indoors under artificial light, make sure the whole page is evenly lit.
  3. Whole page in frame. Get all four corners of the page in the shot. If the page is bound, include enough of the gutter so we can see the text clearly.
  4. Steady hand. Use a tripod or rest the camera on a book or box to keep it still. Blurry photos don't extract well.
  5. Don't retouch. Upload photos as-is, straight from your phone or camera. The system handles EXIF rotation, exposure correction, and sizing automatically.

Formats​

The system accepts:

  • JPEG — the default. Most phone cameras save JPEGs.
  • PNG — lossless, larger file size. Useful if you want to preserve every pixel.
  • WebP — modern, efficient. Works on any modern phone or camera.
  • TIFF — for archival-grade scans. Single-page TIFFs only (we don't support multi-page TIFFs in the photo upload path).
  • HEIC — what an iPhone saves by default. No need to convert first.

If your camera saves in some other format (RAW, for instance), convert to JPEG before uploading.

Uploading​

  1. On the Retroconvert page, under 2 · Source catalogue, drop your photos on the upload area or click to browse. (A PDF goes in the same place — one or the other, not both.)
  2. Photos appear in the order you chose them, one row per page.
  3. Check the order — use Move earlier and Move later to shift a photo, remove any that aren't right, and Add more photos to top the batch up.
  4. Click Use these N pages when you're ready. The system binds them into a single document to read from.

Troubleshooting​

Photos are blurry. Re-photograph under better light, or with the camera more steady (use a tripod).

Text is too small to read. Move closer to the page, or photograph at a higher resolution. The system needs to read the text, not display it at thumbnail size.

Some pages are missing. You can add more photos after choosing the first lot — use Add more photos before you click Use these N pages.

A page is out of sequence. Use Move earlier / Move later on its row to put it right before you click Use these N pages. Orientation is handled for you from the photo's own metadata.

What happens next​

Once you've uploaded the photos, the system assembles them into a single PDF (EXIF-rotated, downscaled to a consistent size, flattened to white background). That PDF is then processed as if it were a scanned document — the retroconversion pipeline stays the same, and every page still links back to its original photo.

Tips for different materials​

Ledgers and registers — these are often large, heavy, or awkwardly bound. A flat surface and good overhead light are key. If the binding gets in the way, photograph the left and right pages separately and reorder them together.

Daybooks — often smaller and easier to handle. Photograph at a slight angle if the binding pulls the pages up.

Damaged pages — if a page is faded, torn, or stained, do your best to capture what's there. The system can handle quite poor scans; it's designed for historical documents. If you're unsure, take two photos (one with flash, one without, for example) and upload both — you can delete the less useful one after reviewing.

Handwriting — the system's OCR is designed for printed text. Handwritten pages extract less reliably but are still supported. The system shows you confidence scores — anything marked uncertain still needs human review, but the review goes faster than retyping.