JPG vs PDF for scanned documents

A JPEG is one picture with one quality setting. A PDF is a page tree that can hold many pictures, plus fonts, plus a MediaBox that tells a printer the paper size. For a single receipt, JPEG is simpler. For a packet a portal will only accept as “application/pdf,” you wrap the pictures.

Why the PDF is often larger than the sum of JPEGs

Sometimes it is not. If you embed twelve 400 KB JPEGs as DCTDecode streams, the PDF is about 4.8 MB plus a few kilobytes of page objects. People are surprised when it is 28 MB. That happens when the scanner app saved PNG (lossless) or 600 dpi TIFF-in-PDF, or when you chose Fit on 12-megapixel phone photos without resizing. A 4032×3024 JPEG at quality 92 is often 2–3.5 MB. Twelve of those is 24–42 MB before the PDF wrapper.

Shared resources go the other way: one 4 MB letterhead JPEG referenced by 40 pages is 4 MB in a PDF, not 160 MB. Split page 1 out and that 4 MB still has to come along. That is why a “one-page extract” can look huge.

When to use which

Email a photo of a receipt to yourself: JPEG, or resize to 1600 px on the long edge first. Submit “scan of passport + form + photo” to a visa site: PDF, Letter or A4 as the site specified, using JPG to PDF. Fit is wrong when they asked for A4; the page box will not match their stamper.

Need searchable text later? JPEG will never be searchable until you OCR. A PDF from a real scanner with an OCR layer will. Phone photos converted here are pictures; run PDF to Word with OCR if you need a string you can edit.