Resolution is measured in dots or pixels per inch (dpi or ppi). A scan at 300 dpi captures 300 pixels along every inch of paper, so an A4 sheet becomes roughly 2,480 by 3,500 pixels. Double the setting and the pixel count grows four times. That is why a stack of high-resolution colour scans can reach tens of megabytes while a portal accepts a few. This guide gathers what official and vendor sources say about the right setting, then shows how to meet a size limit without losing the good original.
What the sources recommend
No single number fits every document. The sources below come from archives, an OCR project and a scanner maker.
- US National Archives (NARA). The Chief Records Officer's blog states: "When producing digital images of modern textual paper records, a minimum resolution of 300 pixels per inch must be used." It adds that agencies "must digitize in an acceptable RGB color mode if records contain color." This is a rule for US federal records, not for every upload form, but it shows what 'good enough to keep' means.
- Tesseract OCR. The documentation of this widely used open-source OCR engine says: "Tesseract works best on images which have a DPI of at least 300 dpi, so it may be beneficial to resize images."
- Epson. Epson's scan resolution guidelines give 96 to 150 dpi for email, viewing on screen or posting on the web, and 200 to 300 dpi to print, fax, convert to editable text (OCR) or create a text searchable PDF. The same page warns that raising the resolution captures more detail but brings "Larger file sizes" and longer scan and transfer times.
- Library of Congress. Its guide to electronic preservation of visual materials gives an example of scanning text at "600 dpi binary or 300 dpi 8 bit grayscale" and describes making an access image derived from the preservation image, so the high-quality master is kept and a lighter copy is produced for use.
Taken together, 300 dpi is the common floor when text must be read, searched or archived. Lower settings such as 150 dpi are described for email and screen viewing, not for documents that need to stay legible when printed or processed by OCR. The last point, scan once and derive a smaller copy, is the idea the rest of this guide follows.
The portal's own rule comes first
A government or university upload form may state its own size, format or resolution rules, and those override any general advice. Many only state a file size limit and say nothing on dpi. See why upload portals reject scanned documents for examples of how rules differ, and read the note on the upload screen before you scan. If a portal gives a limit in kilobytes, the task pages for compressing a PDF to 1 MB and the other sizes show the typical targets.
A practical order of work
The following is judgement based on the sources above, not an official rule.
- Scan at 300 dpi. Use colour only if the document has colour that matters, such as a stamp or a coloured form. Black text on white paper reads well in grayscale.
- Keep the scan. Treat it as the master, in line with the preservation-and-access idea above.
- Make an upload copy from the master. Reduce the size until it fits the portal, and check that the text is still readable at full zoom.
- If searchable text matters, run OCR on the copy you will keep for yourself, not on a copy squeezed to the smallest size.
How to make a smaller copy with DocPro
Compress PDF runs in your browser and the file is not uploaded.
- Open Compress PDF and choose the scanned PDF.
- Under Quality, pick a level. High quality is 144 dpi with sharp text and the smallest saving. Balanced is 108 dpi and is described as good for email. Small is 72 dpi and readable on screen. Smallest is 54 dpi and is described as a last resort.
- Or tick Aim for a target size and enter a Target size (KB). The tool lowers quality first, then resolution, until the file fits. If it cannot reach the target, it tells you the closest size it reached.
- Press Compress PDF and compare Before and After. Zoom in on the result to confirm that small print is still readable.
Compression re-saves each page as a picture, so selectable text and links are lost. If the compressed copy must stay searchable, run OCR to PDF on the result and press Make searchable PDF. The OCR tool's own advice is to scan straight and in good light at around 200 to 300 dpi.
Other options for a file that is still too large:
- Grayscale converts colour pages to grayscale or black and white, with Resolution choices of 100, 150 or 200 dpi. This often shrinks a colour scan a lot when colour does not matter.
- Split PDF cuts a long scan into parts by Every N pages or By ranges when a portal accepts several files. It does not split by file size.
- Images to PDF builds one PDF from phone photos of pages. See the task page for scan to PDF. Very large photos make a large PDF, so run the result through Compress PDF if needed.
Common mistakes
- Scanning at a low setting to save space. A 100 dpi scan cannot be improved later. Compressing a 300 dpi master gives a better result than rescanning at a low dpi.
- Using the highest setting for everything. The Library of Congress example of 600 dpi is for binary scans in a preservation context. For ordinary letters and forms it mostly adds file size.
- Enlarging a small scan. Resizing a 150 dpi image to 300 dpi does not add detail. OCR and readers see the same blur.
- Running OCR and then compressing. Compression turns pages back into pictures and drops the text layer, so run OCR on the compressed result, not before it.
What DocPro cannot do
DocPro does not control your scanner and cannot raise the real resolution of a scan that was captured at a low setting. Compress PDF lowers resolution and quality, so very small targets on long documents may not be reachable. If a portal requires a specific dpi value, check how it measures it and test with a sample, because a PDF has no single resolution setting once pages have been re-saved.
Frequently asked questions
Is 200 dpi enough for a PDF upload?
Epson's guidance lists 200 to 300 dpi for OCR and text-searchable PDFs, while NARA requires at least 300 ppi for modern paper records and Tesseract works best at 300 dpi or more. 200 dpi can read well for clean printed pages, but check the result at full zoom and follow the portal's own rule.
Is 600 dpi better than 300 dpi?
Not for most letters and forms. It captures more detail and produces much larger files. The Library of Congress example pairs 600 dpi with binary scans for preservation; for ordinary documents 300 dpi is the usual floor.
Should I scan in colour or grayscale?
NARA says to use an acceptable RGB colour mode if records contain colour. For black text on white paper, grayscale is usually enough and makes smaller files. Keep colour for stamps, seals and coloured forms.
Why is my scanned PDF so large?
Each page is a picture, and the pixel count grows with the square of the dpi. Colour and high dpi together make the biggest files. Compress PDF or Grayscale can shrink them.
Is my scan uploaded when I compress it?
No. Compress PDF runs in your browser and the file stays on your device.