How to compress a scanned PDF and keep it readable

Start by identifying the page content

A PDF can contain selectable text, page images, or a mixture of both. A scan often behaves like a photograph when you try to select words. Some scans also have an OCR text layer, so selectable text alone does not prove that the page is made from ordinary text objects.

Look at several pages rather than judging the entire document by its cover. A file can mix a digital introduction with hundreds of scanned pages. Colour, photographs, large dimensions, and high-resolution images can all contribute to size.

Match the compression attempt to the document

High Quality uses lossless optimization. Smart Compression and Maximum Reduction can reduce supported image content, with the stronger setting accepting a greater image-quality tradeoff. The output depends on the encoding and existing compression of the images, so equal-looking mode results can still be legitimate for some documents.

For a scan that must fit a submission limit, try a target such as 5 MB, 10 MB, or 20 MB and review the actual result. A long or highly detailed scan may not meet that target without unacceptable damage. The maximum input is 300 MB on supported desktops. Device-specific limits still apply, and some files may retain a lossless result when image processing exceeds resource or validation limits.

Inspect the details that compression can hide

Compare at least an ordinary text page, the smallest print, a page with a stamp or signature, and any important diagram. If the document contains coloured marks that carry meaning, check them as well. A readable heading is not enough to establish that an entire scan survived well.

Do not confuse compression with OCR. Compression aims to reduce bytes. OCR tries to recognize text from page images and may be useful if you need an editable Word document. Neither process can reliably reconstruct detail that was missing from the original scan.

• Check names and identification numbers character by character.
• Inspect thin lines, decimal points, and pale handwriting.
• Keep the original scan in case a reviewer asks for better detail.

Handle large scans with the device in mind

A browser may need temporary storage and extra working memory beyond the size of the selected file. The accepted size shown in the upload area is a limit for that device, not a promise that every complex document below it will finish equally quickly.

Use a capable desktop for demanding documents, keep the processing tab open, and avoid running several large conversions simultaneously. If the source is unreadable or the result cannot meet the required size, splitting permitted sections or rescanning at a suitable resolution can be more useful than repeating the same attempt.