PDF compression guides
How to Compress a Scanned PDF Without Making Text Hard to Read
Treat readability as something to inspect, not something a compression percentage can prove.
Last updated
“Without losing readability” is a verification goal, not an absolute promise. Scanned PDFs store pages as images, so reducing resolution, JPEG quality, or color information can remove detail. Start with the least aggressive setting that meets the size target, then inspect the export at 100% and at the larger zoom recipients may use. Keep the original, and never use a compressed access copy as the preservation master for archival, legal, identity, or print-production work.
First decide whether the pages are scans or digital content
Try selecting a sentence and searching for a word you can see. Selection usually points to digital text or an OCR text layer; a page that behaves like one large picture is scan-heavy. Mixed PDFs need page-by-page judgment because a text cover can hide image-only attachments later in the file.
| What you find | Typical signal | Safer starting point |
|---|---|---|
| Mostly digital PDF | Text selects cleanly and stays sharp at high zoom. | Use Smart compression so text remains text where possible. |
| Scan-heavy PDF | Each page acts like an image; dust, skew, or paper texture is visible. | Use Scanned PDF compression with Clarity first or Balanced. |
| Mixed PDF | Some pages select as text while forms, photos, or appendices do not. | Try Smart compression first, then inspect every image-based page. |
Understand what scan compression changes
| Change | Why size may fall | What can become unreliable | Restrained use |
|---|---|---|---|
| Lower image resolution or downsample | Fewer pixels describe each page. | Fine print, thin strokes, barcode edges, and line art can soften. | Reduce gradually and inspect at 100% plus the zoom level recipients will use.1 |
| Convert meaningful color to grayscale | Each pixel carries less color information. | Highlights, stamps, annotations, and color-coded diagrams may lose meaning. | Turn off Keep scans in color only when color is not evidence or content.3 |
| Increase lossy image compression | More pixel detail is discarded. | Halos, block artifacts, blurred handwriting, and damaged photo detail can appear. | Stop at the first setting that meets the practical target.1 |
| Resample each page into a new image-based PDF | A consistent image format replaces mixed page content. | Selectable or searchable text can disappear, and existing signatures or interactive elements may not survive. | Use Scanned PDF compression on an expendable copy, not the only original. |
The current FADGI technical guidelines give material-specific capture and quality-control requirements rather than one universal web-PDF recipe. For modern textual records, the guideline describes 300 ppi-class capture, color or grayscale, approved master formats, and visual evaluation when compression is used.3
Read the scan fixture as an example, not a forecast
In PDFStay's fixed scan-style fixture, the output fell from 2,141,091 bytes to 187,107 bytes, a 91.26% reduction, while the recorded page count went from 2 to 2 and the automated readability field was true. The methodology page and JSON attach a strict boundary: this one browser run is not a promised rate or visual-quality score for another scan.4, 5
| Fixture | Browser version | Input bytes | Output bytes | Reduction | Input pages | Output pages | Automated readable-PDF check | Boundary |
|---|---|---|---|---|---|---|---|---|
| scan-style-pdf-v1 | 148.0.7778.96 | 2,141,091 bytes | 187,107 bytes | 91.26% | 2 | 2 | true | Fixed PDFStay fixture; not a promised compression rate or visual-quality score |
Fixed PDFStay fixture; not a promised compression rate or visual-quality score
Inspect the details an automated size check cannot approve
FADGI separates technical validation from visual inspection and calls for viewing image detail at 100% magnification during quality control. For a personal PDF, inspect every page when the document is short and sample repeated page types only after confirming completeness.3
- Fine print and low-contrast text: compare small dates, decimal points, footnotes, and faint carbon-copy marks.
- Signatures and handwriting: check pen strokes, initials, dates, and any pressure-light marks.
- Stamps and seals: confirm edges, serial numbers, embossing shadows, and ink colors that carry meaning.
- Barcodes and QR codes: scan them from the exported PDF, not from the original.
- Diagrams and forms: look for thin rules, arrowheads, legends, checkbox states, and alignment.
- Photos and halftones: compare faces, identification details, gradients, and dark areas for artifacts.
- Page count and order: match the output count to the input and move through the first, middle, and last pages.
Keep the original and avoid repeated lossy recompression
Save the first-generation scan as the master and give the compressed copy a new, descriptive name. If the copy is still too large, return to the master and make one revised export. Recompressing an already lossy copy stacks permanent pixel changes and makes it harder to identify which pass damaged a detail.1
Know when compression needs a specialist check
| Document role | Do not assume | Required check |
|---|---|---|
| Archival or preservation master | A readable access copy satisfies the archive's capture, format, metadata, or compression specification. | Retain the master and follow the current archive or records authority requirements.3, 2 |
| Legal or digitally signed PDF | A visually similar export preserves evidentiary status, form behavior, or signature validity. | Keep the signed original and validate any working copy in the required signature or case-management software. |
| Identity document or application evidence | A face, number, stamp, or security feature is acceptable because the page still opens. | Follow the recipient's upload rules and inspect every character and visual identifier. |
| Print-production PDF | A smaller screen copy preserves image resolution, embedded fonts, layers, transparency, and output intent. | Ask the printer for a preflight profile and approve a proof before replacing production artwork.1 |
FAQ
Can scan compression guarantee no quality loss?
No. Lossless encoding exists, but substantial size reduction often involves downsampling or lossy image compression. Adobe notes that image optimization can remove pixel data permanently, so readability and visual detail must be inspected.1
Which PDFStay scan setting should I try first?
Use Clarity first when small writing or marks matter, or Balanced for an ordinary sharing copy. Move to Smaller file only when the target is strict, then repeat the full visual checklist.
Why might text stop being selectable after scan compression?
Scanned PDF compression can rebuild pages as images. The letters can remain visible while the underlying text objects or OCR layer are no longer present, so test selection and search separately from visual readability.
Sources
- Adobe Acrobat — PDF Optimizer settingsofficial technical documentationSource reviewed:
- U.S. National Archives — Scanned Images of Textual Recordsgovernment technical guidanceSource reviewed:
- FADGI — Technical Guidelines for Digitizing Cultural Heritage Materialsgovernment technical guidanceSource reviewed:
- PDFStay reproducible compression checksPDFStay first-party methodologySource reviewed:
- PDFStay compression benchmark data (JSON)PDFStay first-party dataSource reviewed: