Photos and Scans to PDF: Page Size, DPI and Ordering Done Right

Every PDF page has an explicit size in physical units — A4 is 210 by 297 millimetres, US Letter is 8.5 by 11 inches — and each image has to be placed inside that box. There are only two honest options: fit the whole image inside the page, leaving margins in the empty dimension, or fill the page and crop the overflow. A converter that silently stretches the image to fill the page distorts it, which is the failure people notice last and hate most.

Page size should follow the destination. A4 for most of the world, Letter for North America; both for a mixed audience is a bad idea because a document with alternating page sizes prints badly and looks unprofessional. If your images are a mixture of portrait and landscape, decide whether to keep each photo's orientation as its own page orientation, or to force one orientation and accept margins on half the pages.

A PDF stores an image in pixels and scales it to a physical size, so the same photo can be 100 DPI or 400 DPI depending on the page it is placed on. Take a 4000-pixel-wide photo and put it across an A4 page: 4000 divided by 8.27 inches gives about 484 dots per inch. That is well above what any office printer resolves. Put the same photo across a 4-inch-wide region and it is 1000 DPI, which is pure waste.

For print, 300 DPI at the final placed size is the standard target, and 150 DPI is acceptable for internal documents. For screen-only PDFs, 100 to 150 DPI is enough and keeps the file small. The important asymmetry is that you cannot add resolution by asking for it: telling a converter to write a 2000-pixel image at 600 DPI enlarges it to a blurry 600 DPI, it does not create detail. Downsizing, on the other hand, genuinely saves space.

Page order comes from filename order, and filenames sort as text: page1, page10, page2. Zero-pad the numbers — page01, page02, page10 — or the PDF will be in an order that only makes sense to a computer. Sorting by modification date instead helps when the files came from a camera roll in sequence.

Rotation is the second trap. A photo taken sideways is stored with an EXIF orientation flag, and if the converter does not apply it, that page arrives rotated inside the PDF, where you can no longer fix it by rotating your viewer. Check the first and last page of the output before sending it anywhere. Also watch for blank or duplicated pages: sheet-feed scanners occasionally pull two sheets at once, and a phone scan of a two-sided document produces one page per side whether you wanted that or not.

An image-only PDF is roughly the sum of the JPEGs inside it, so size is driven by pixel count and quality setting. Recompressing at around 85 percent quality is usually visually indistinguishable from the original photo and cuts the size substantially. Downsampling anything above 300 DPI to 300 DPI saves more, and converting text-only scans to grayscale can halve them again.

Never convert a PDF page back to an image and then into a new PDF to 'compress' it — that is two lossy generations for a file that will look worse and often end up larger. If the document needs to be searchable or selectable, note that a photo of text is just a picture: making the text real requires optical character recognition, which is a separate step from assembling the PDF.