How to convert JPG images to PDF

Auto gives each page the dimensions of its image, converted at 72 points per inch. No part of the image is hidden and no empty space is added, and a mixed set of portrait and landscape photos simply produces mixed pages. This is the right default for a photo album, a set of receipts, or anything that will only ever be viewed on screen.

A standard size is what you want when the document will be printed or filed. A printer feeds sheets of a fixed size, and pages that vary get scaled inconsistently or rejected. A4 is 595 × 842 points and Letter is 612 × 792, and choosing one means every page in the output matches it regardless of what the images were.

The moment you fix the size, the aspect-ratio mismatch appears and the fit setting starts to matter. That is the trade: auto guarantees fidelity and irregular pages, fixed guarantees consistent pages and requires a decision about the mismatch.

Contain scales the image until it fits entirely within the page, leaving bands of empty page along two edges when the ratios differ. Nothing is lost and nothing is distorted. For a scanned document this is nearly always correct, because a cropped edge might be a page number, a signature line or a margin note.

Cover scales until the image fills the page completely and lets the overflow fall outside. The page looks deliberate and edge-to-edge, at the cost of losing image content on two sides. Fine for a photograph where the subject is central, wrong for anything where the edges carry meaning.

Fill stretches the image to the page in both directions independently, ignoring the aspect ratio. Everything is shown, nothing is cropped, and the result is visibly distorted — faces widen, circles become ovals, text leans. It exists for the rare case where the image was already the right proportions and you want to force an exact match, and it is almost never the answer otherwise.

The image is embedded as-is. A JPEG goes in carrying its existing compressed data, so it is not re-encoded and does not pick up another generation of loss. The PDF is therefore roughly the sum of your image files plus a small structural overhead, and a set of 4 MB phone photos makes a large PDF — the wrapping does not compress them.

What does not carry through is anything requiring a text layer. A photographed page of text is a picture of text: the PDF is unsearchable, and it stays unsearchable until OCR adds a text layer. If you are converting scans specifically so they can be filed and found later, plan on that second step.

One detail worth checking on phone photos: images often carry an EXIF orientation flag rather than storing rotated pixels. If a photo appears sideways in the output, the flag is the likely cause, and the rotation tool fixes the affected pages after the fact.