Understanding image-to-PDF
Photos in, paper-shaped file out.
The mismatched logic of pixel images and page-based documents — and how to make them coexist.
PDF is page-shaped.
A PDF file is a collection of pages, each with a fixed size (A4, US Letter, custom). An image is a flexible rectangle of pixels. To put one inside the other you have to decide: how big should the page be, where on it does the image sit, and how should the image scale. Get any of those wrong and the output looks awkward — a tiny photo floating in white, or a panorama cropped to a square.
image (pixels) → page (millimetres) + DPI
Pages, photos and DPI.
A 4032×3024 photo from a phone camera is 4032 pixels wide. On A4 paper (210 mm wide) at 300 DPI you'd want 2480 pixels — so the photo is comfortably big enough to print sharply. At 600 DPI you'd want twice that, and the same photo would start to look soft. The DPI you choose tells the page how to map pixels to millimetres. For screen-only PDFs (most email attachments), DPI is a fiction; for printable scans, it matters.
Fit, fill, or letterbox.
When the photo's aspect ratio doesn't match the page's, you have three options. Fit shrinks the photo to fit inside the page, leaving margins where the ratio doesn't match. Fill scales until it covers the page, cropping the overhang. Stretch distorts the image to match and is almost always wrong. Receipts, ID copies and scanned documents nearly always want fit; photo books often want fill.
One page per image, or many.
The default is one image per page — a stack of sheets, each showing a single photo at the page's natural orientation. Some workflows want a contact sheet (multiple photos per page in a grid); some want a strip of receipts on a single page; some want each photo printed on its own auto-sized page. Pick the model first; the bytes follow.
Embedded vs re-encoded.
The most efficient way to put a JPG in a PDF is to embed the original bytes — PDF supports JPEG natively, no re-encoding required, no quality loss. PNGs and HEICs need conversion (PDF doesn't speak either), which means decoding and re-encoding to a flat raster or to a JPEG2000 stream. The output PDF is therefore best-quality when the source was JPG, and slightly larger but still good when it was PNG. HEIC always needs a conversion step.
Orientation and rotation.
A phone photo carries an EXIF orientation tag. Embed the file naïvely and the PDF reader will display it in the original sensor orientation — sometimes sideways. The right fix is to read the EXIF orientation, rotate the pixels (or rotate the PDF page), and clear the tag. Tools that get this wrong produce PDFs full of sideways photos; tools that get it right just work.
Read next