How OCR Affects a Scanned PDF's Size and How to Keep It Small
People often run OCR on a scanned document and notice the file got bigger, then assume OCR is the culprit. It almost never is. Understanding what OCR actually adds, and what really drives a scanned PDF's size, lets you keep a searchable document small without giving up the text layer that makes it useful.
The situation
You scanned a stack of pages, ran OCR so the text becomes searchable and selectable, and the resulting PDF is large, maybe larger than before. You need to email it, upload it to a portal, or store many of them, and the size is a problem. The instinct is to skip OCR to save space. That is the wrong lever, because the text OCR adds is tiny compared with what is really making the file heavy.
Why OCR barely changes the size (the mechanism)
OCR does not replace your scanned pages. It reads the image of each page and adds an invisible text layer on top, positioned behind the image so you can select and search the words while still seeing the original scan. That text layer is just characters and coordinates. Compared with the page images underneath it, the invisible text layer is usually very small, often a negligible fraction of the file.
So what makes a scanned PDF big? The images. Each page is stored as a picture, and the size of those pictures is driven by three things:
- Resolution (DPI): a 600 DPI scan holds four times the pixels of a 300 DPI scan of the same page.
- Color depth: full color pages are far heavier than grayscale or black-and-white.
- Image encoding: how efficiently each page image is compressed.
If your OCR'd PDF grew, it is usually because the OCR step re-saved the pages at a high resolution or in color, or added a second copy of the image, not because the text layer is heavy. The fix is to control the images, not to drop OCR.
The OCR-plus-downsample workflow
The reliable way to get a small, searchable PDF is to treat OCR and image reduction as two separate jobs and do both.
- Scan at a sensible resolution. For text documents, 300 DPI is plenty for OCR accuracy. Scanning at 600 DPI quadruples the data for little gain in recognition.
- Run OCR to add the text layer. This makes the document searchable and copy-paste friendly while keeping the original page image visible.
- Downsample the page images. Reduce the images to a readable on-screen resolution such as 150 DPI, which is where most of the saving comes from. This does not remove the text layer; it only shrinks the pictures behind it.
- Convert color to grayscale if color is not needed. For a plain text scan, grayscale or bitonal removes a large share of image data.
- Compress to a target. Finish by compressing the file to the size you actually need to send or store.
Done in this order, you keep every searchable word and still land at a fraction of the original size.
Searchable PDF versus PDF/A
Two output choices often come up:
- Searchable PDF is the everyday result: your scan plus the invisible text layer. It is the smaller, more flexible option and is what you want for email, uploads, and general use.
- PDF/A is an archival format designed for long-term preservation. It embeds everything needed to reproduce the document later, including all fonts, which can make the file larger. Choose it when a rule or archive requires PDF/A; otherwise a plain searchable PDF is lighter.
If you are optimizing for size, a searchable PDF is the better default. Reserve PDF/A for cases where archiving standards demand it.
Once your searchable PDF is ready, PDF300 lets you enter a target file size and compresses to fit it, so you can shrink the scan images while the text layer rides along untouched. Small files are re-encoded in the browser and never leave your device; a server mode is available for higher-quality results over an encrypted connection.
Checklist
- [ ] Scanned text at about 300 DPI rather than an excessive resolution.
- [ ] Ran OCR to add the searchable text layer.
- [ ] Downsampled the page images to a readable on-screen DPI.
- [ ] Converted to grayscale or bitonal when color was not needed.
- [ ] Chose searchable PDF over PDF/A unless archiving rules require it.
- [ ] Compressed the final file to your target size.
FAQ
Does OCR make a PDF bigger?
Only slightly, if at all. OCR adds an invisible text layer that is usually tiny next to the page images. If your file grew noticeably, it is almost always because the OCR step re-saved the images at a higher resolution or in color, not because of the text itself.
Can I keep a PDF searchable and still make it small?
Yes. Downsampling the page images and converting them to grayscale shrinks the file without touching the invisible text layer, so the document stays fully searchable. Do the image reduction after OCR, then compress to your target size.
Should I use searchable PDF or PDF/A to save space?
Use a plain searchable PDF when size matters, because it is lighter and more flexible. PDF/A embeds everything for long-term preservation, including all fonts, which tends to make files larger. Choose PDF/A only when an archive or regulation specifically requires it.