Poor Text Quality in Scanned PDFs: Diagnosis and Fixes

Before applying a fix, identify what's actually wrong. Poor scanned PDF quality typically falls into one of these categories: **Scan resolution too low**: If the scan was done at 72 DPI or even 100 DPI (web resolutions), the text will look blurry when zoomed in, and OCR will struggle with fine letter details. Text requires at least 200 DPI, and 300 DPI is the minimum professional standard. **Over-aggressive JPEG compression**: Many scanners apply JPEG compression to reduce file size. Too much compression creates visible artifacts — blotchy patches, halos around letters, loss of fine detail in thin strokes. This makes OCR particularly difficult. **Poor scan conditions**: Physical issues — misaligned document, dirty scanner glass, wrinkled paper, shadow from page curl, low light — create problems that no amount of post-processing can fully fix. **Grayscale compressed to black-and-white**: Some scanners or scan-to-email features automatically convert to black-and-white or bilevel images. While this produces tiny file sizes, it loses subtle gray tones that help distinguish letters, especially in older documents or documents with decorative fonts. **OCR errors in text layer**: Some scanners add an OCR text layer automatically. If this OCR was inaccurate, the underlying image may look fine but the text you can select or copy is wrong.

If you're doing the scanning yourself and have control over settings, these are the knobs to get right.

1. 1Set resolution to at least 300 DPI for standard documents, 400 DPI for documents with small text or fine details
2. 2Use grayscale mode rather than black-and-white (bilevel) for most documents — grayscale preserves subtle contrast that helps OCR accuracy
3. 3Use color mode only if the document has meaningful color content — it produces larger files with no quality benefit for text
4. 4Set JPEG quality to 80% or higher, or use PNG/TIFF format if the scanner allows it (these are lossless)
5. 5Clean the scanner glass before scanning and press the document flat to avoid page curl shadows
6. 6Perform a test scan of one page and zoom in at 100% before scanning the whole document — verify text is sharp and clear

Even with imperfect source scans, OCR can often be improved through preprocessing. Before running OCR, consider these enhancements: **Deskewing**: If pages were scanned at a slight angle, text rows are diagonal rather than horizontal. OCR engines work best on straight text. Deskewing corrects the angle and dramatically improves recognition accuracy. **Contrast enhancement**: Increasing contrast in a low-contrast scan (old yellowed documents, faded photocopies) makes text and background more distinct, which helps OCR. **Noise reduction**: Scanner noise (random specks and dots in the background) confuses OCR engines into treating noise pixels as letter fragments. A mild blur or median filter reduces noise while preserving letter shapes. **Binarization threshold**: For black-and-white conversion, the threshold that separates 'text' from 'background' matters enormously. Wrong threshold means text becomes incomplete or background noise becomes text. LazyPDF's OCR tool uses Tesseract, a robust open-source OCR engine that handles a wide range of document quality levels. For best results, upload a clean scan at 300 DPI or higher. The tool will extract text from the images and make the document searchable.

1. 1Upload your scanned PDF to LazyPDF's OCR tool
2. 2Select the correct language for the document — OCR accuracy depends heavily on the language model
3. 3Process and download the OCR'd PDF
4. 4Open the result and test text selection by clicking and dragging — if text selects correctly, OCR worked
5. 5Search for a specific word (Ctrl+F or Cmd+F) to verify the text layer is accurate

Sometimes you're stuck with a poor scan and can't redo it. Here are strategies for working with low-quality scanned PDFs: **Read the image, not the text**: For personal use where you just need to read the document, zoom in to 200-300% to read blurry text more easily. Your eye is much better at interpreting blurry letter shapes than OCR software. **Manual transcription for critical content**: If you need accurate text from a poor scan, manual typing may be necessary. OCR errors in important documents (contracts, court documents, medical records) can have serious consequences. **Use multiple OCR tools**: Different OCR engines have different strengths. If Tesseract struggles with a scan, Google Docs' free OCR (upload PDF, open with Google Docs) or Microsoft OneNote's OCR may handle it differently and get more accurate results. **Enhance before compressing**: If you need to compress a scanned PDF (to reduce file size), compress after any OCR and enhancement steps — not before. Starting from the highest quality scan preserves options.

High-resolution scans produce large files. A 300 DPI scan of a 20-page document can easily be 50-100 MB. Compressing such a file is legitimate — but the compression needs to be done carefully to avoid destroying the quality you worked to create. LazyPDF's compress tool uses Ghostscript settings that downsample images to appropriate resolutions for their intended use. 'eBook' quality settings target 150 DPI — adequate for screen reading. 'Screen' quality goes lower. 'Printer' settings preserve higher resolution. For scanned documents, use 'eBook' or 'Printer' quality compression — never 'Screen' quality if you need readable text. The size difference between 'eBook' and 'Screen' compression is often not worth the quality loss for scanned documents with thin text strokes.