Scanned PDF Too Large? Reduce DPI + Convert to JPEG in Practice

Bottom line first: Scanned PDFs are too large because high-DPI scanning and uncompressed bitmap formats account for over 90% of the file size. The solution is a 3-step approach: reduce DPI to 150, convert BMP/lossless formats to JPEG at quality 75, and add an OCR text layer. A 120-page color contract scan can be compressed from 95MB to 3.2MB, with text remaining clearly legible, searchable, and copyable. Below, we start by examining what makes files large, then provide detailed parameters for each of the 3 steps, covering differentiated strategies for black-and-white, grayscale, and color scans.

If you're not yet familiar with the overall principles of PDF compression, we recommend reading PDF Compression Principles and Methods Explained.

1. Why Are Scanned PDFs So Large?

Scanned PDFs embed bitmap images generated by the scanner page by page into a PDF container. Unlike electronically generated PDFs, each page of a scanned PDF is a complete image, so the file size is entirely determined by the images. Understanding the sources of file size is key to targeted compression.

Size Source	Typical Share	Cause	Compression Potential
High-DPI scanning	50%–65%	300/600DPI scanning, far exceeding reading needs	★★★★★
Lossless bitmap format	20%–35%	BMP/uncompressed TIFF/FlateDecode, no lossy compression	★★★★★
Color mode waste	5%–15%	Pure text using 24-bit color instead of 1-bit black and white	★★★★☆
Multi-page accumulation	5%–10%	Many pages, each adding redundancy	★★★☆☆
PDF structure overhead	1%–3%	Page objects, cross-reference tables, etc.	★☆☆☆☆

As the table shows, high DPI and lossless formats together account for over 80%, making them the primary targets for compression. A 120-page 300DPI color scan has about 0.8MB per page, totaling 95MB — but screen reading only needs 150DPI, and after converting to JPEG, each page can be reduced to 0.027MB, totaling just 3.2MB.

2. 3-Step Compression Method in Detail

Based on the sources of scanned PDF file size, there are 3 standardized compression methods. Each step has different principles, applicable scenarios, and effects. The table below provides an overview, followed by detailed explanations.

Step	Method	Principle	Expected Compression	Clarity Impact
1	Reduce DPI	Lower pixel density	50%–75%	Negligible for screen reading
2	Format conversion	BMP/TIFF→JPEG	80%–95%	Quality 75 acceptable to the eye
3	OCR text layer	Generate searchable text	+5%–10% size	Improves usability

1. Reduce DPI (Highest ROI)

DPI (dots per inch) determines the pixel density of a scanned document. Scanners default to 300DPI or even 600DPI, but screen reading only requires 96–150DPI, and printing only needs 200–300DPI. Reducing 600DPI to 150DPI cuts the pixel count to 1/16 of the original, directly reducing size by 94%. Even reducing to 200DPI achieves an 89% reduction.

In practice: open the scanned PDF with SmartSlim, and the Rust compression engine analyzes images page by page, automatically selecting the optimal DPI based on scan type — pure text black-and-white scans are reduced to 200DPI, grayscale scans to 150DPI, and color scans to 150DPI. Downsampling uses the Lanczos algorithm to maintain sharp edges, avoiding the blurring that simple scaling causes.

Scan Type	Original DPI	Recommended DPI	Pixel Reduction	Readability
B&W text	300	200	56%	Clearly legible
B&W text	600	200	89%	Clearly legible
Grayscale text	300	150	75%	Clearly legible
Color document	300	150	75%	Slightly faded, acceptable
Color photo page	600	200	89%	Slight detail loss

For a detailed explanation of the difference between DPI and PPI, see DPI vs PPI: Scanning and Printing Resolution Explained.

2. Format Conversion (Largest Reduction)

Images in scanned PDFs commonly use three encodings: FlateDecode (PNG's lossless compression), CCITT G4 (the black-and-white fax standard), and DCTDecode (JPEG). Many scanners default to FlateDecode or no compression at all (BMP), resulting in huge file sizes. Converting to JPEG at quality 75 can cut 80%–95% of the size.

For black-and-white scans, there's an even better option — JBIG2 encoding. JBIG2 is designed specifically for black-and-white text, using template matching to identify repeated glyphs, making it 50%–70% smaller than CCITT G4. A 100-page black-and-white contract is about 40MB with FlateDecode, 8MB with CCITT G4, and only 2MB with JBIG2. SmartSlim automatically detects black-and-white scans and enables JBIG2 encoding.

Encoding	Applicable Type	100-page Size	vs. FlateDecode
FlateDecode (lossless)	General	40MB	Baseline
CCITT G4	B&W	8MB	-80%
JBIG2	B&W	2MB	-95%
JPEG q75	Grayscale/Color	3.5MB	-91%
JPEG q85	Grayscale/Color	5.2MB	-87%

3. OCR Text Layer (Improves Usability)

Scanned PDFs are images by default — not searchable and not copyable. OCR (Optical Character Recognition) overlays a transparent text layer on top of the image, making the scan a searchable and copyable document. The OCR text layer itself only adds 5%–10% to the file size but dramatically improves usability.

An additional benefit of OCR is assisted compression — with a text layer, the image layer can be reduced to an even lower DPI (such as 100DPI), because even if the image is blurry, text can still be retrieved and copied through the text layer. SmartSlim integrates OCR capabilities, automatically generating a text layer during compression, supporting mixed Chinese-English recognition with over 95% accuracy.

3. Case Study: 120-Page Contract Scan from 95MB to 3.2MB

This is a corporate procurement contract in .pdf format, originally 95.4MB. 120 pages of A4, 300DPI color scan, using FlateDecode lossless encoding. It needs to be uploaded to an e-signing platform (10MB limit), with a target of under 5MB.

Document profile: 120 pages A4 color scan, 300DPI (2480×3508 pixels per page), FlateDecode encoding, ~0.8MB per page, no OCR text layer.

Parameters and size changes:

Step	Operation	Key Parameters	Size Change
1	Reduce DPI	300DPI→150DPI, Lanczos downsampling	95.4→28.6MB
2	Color to grayscale	Contract is pure text, 24-bit color→8-bit grayscale	28.6→12.4MB
3	JPEG encoding	FlateDecode→JPEG q75	12.4→3.0MB
4	OCR text layer	Chinese-English recognition, overlay transparent text	3.0→3.2MB

Result: 3.2MB, compression ratio ~29.8:1, 96.6% size reduction. Well below the platform's 10MB limit. Text is clearly legible on screen, 150DPI print quality is good, full-text keyword search is available, and clause text can be copied. The entire process completed locally — contract content never passed through any third-party server, meeting the confidentiality requirements for commercial contracts.

4. Differentiated Strategies for B&W, Grayscale, and Color Scans

Different color modes of scanned documents require different optimal compression strategies. The table below provides comparison recommendations for three types of scans.

Scan Type	Typical Scenario	Recommended DPI	Recommended Encoding	Expected Compression
B&W (1-bit)	Pure text contracts/documents	200	JBIG2	95%–98%
Grayscale (8-bit)	With signatures/handwriting	150	JPEG q75	90%–95%
Color (24-bit)	Color documents/photo pages	150	JPEG q72	85%–93%
Mixed	Contract + color attachment pages	Per-page processing	Per-page selection	90%–96%

A general principle: for pure text, use black-and-white + JBIG2 first (highest compression rate); for documents with signatures, use grayscale + JPEG; for pure color pages, use color + JPEG. SmartSlim analyzes color complexity page by page and automatically selects the optimal encoding — text pages convert to black-and-white JBIG2, signature pages to grayscale JPEG, and color pages to color JPEG — avoiding one-size-fits-all approaches. If you're also having trouble emailing scanned documents, see PDF Too Large to Email? Compress to Under 5MB in 3 Steps.

5. FAQ

Q1: How can I compress a scanned PDF to under 5MB?

Follow the 3-step process: reduce DPI to 150, convert BMP/lossless formats to JPEG at quality 75, and add an OCR text layer. A 120-page color contract scan can be compressed from 95MB to 3.2MB, with text remaining clearly legible, searchable, and copyable. The key is to reduce DPI first (largest share), then convert to JPEG format, and finally add an OCR text layer to improve usability. Black-and-white scans can further use JBIG2 encoding for even smaller files.

Q2: What if text becomes illegible after reducing the scanned PDF's DPI?

For pure text scans, 150DPI is sufficient for clear reading and printing, and even 100DPI remains legible. If text is illegible, it's usually because the original scan was blurry or the DPI was reduced too far. We recommend 200DPI for black-and-white text, 150DPI for grayscale text, and 150DPI for color scans. Additionally, a sharpening filter can enhance text edges. SmartSlim automatically identifies scan types and assigns DPI accordingly, avoiding blurriness caused by one-size-fits-all approaches.

Q3: Are the compression methods the same for black-and-white and color scans?

No. For black-and-white scans, we recommend JBIG2 encoding (50%–70% smaller than CCITT G4), with DPI reducible to 200 while remaining clear. For grayscale scans, convert to JPEG at quality 75 with 150DPI. For color scans, convert to JPEG at quality 72 with 150DPI — the color information loss is acceptable. Black-and-white scans have the highest compression rate (up to 98%), while color scans achieve about 90%–95%. SmartSlim automatically selects the optimal encoding based on the scan's color mode.

Q4: What are the benefits of adding an OCR text layer to scanned PDFs?

An OCR text layer transforms a scanned document from an image into a searchable, copyable document. The benefits are threefold: first, full-text keyword search without flipping through pages; second, text can be copied to other documents; third, assisted compression — with a text layer, the image layer can be reduced to a lower DPI while content remains retrievable through the text layer. The OCR text layer itself only adds 5%–10% to the file size but dramatically improves document usability. SmartSlim integrates OCR capabilities, automatically generating a text layer during compression.

Summary

Scanned PDFs are too large because high-DPI scanning and lossless bitmap formats account for over 80% of the file size. The solution is 3 steps of targeted compression: reducing DPI has the highest ROI (300→150DPI, 50%–75% reduction), format conversion has the largest reduction (BMP→JPEG q75, 80%–95% reduction; black-and-white with JBIG2 can reach 98%), and the OCR text layer improves usability (+5%–10% size for searchability and copyability). With all 3 steps combined, a 120-page color contract can be reliably compressed from 95MB to 3.2MB.

Remember three points: first, select encoding by color mode — black-and-white uses JBIG2, grayscale/color uses JPEG — don't use one approach for all. Second, 150DPI is the golden balance point for scan compression, sufficient for both screen reading and printing. Third, use local compression for sensitive documents like contracts — never upload them to online tools. With the right tools and methods, slimming down scanned PDFs is actually quite simple.