PDF compression guides
Why Is My PDF So Large? 7 Causes and the Fix for Each
Diagnose the dominant content before choosing a PDFStay mode or returning to the source application.
Last updated
Do not start by moving a compression control to its strongest setting. First diagnose what occupies the file: oversized images, image-only scans, unnecessary color data, fonts, redundant objects, complex artwork, or content that has already been compressed. The right fix depends on the signal. Some files need Image optimization or Scanned PDF compression; others need a cleaner export from the source application or should remain unchanged.
Run a five-minute diagnosis before compressing
- Record the byte size and page count, then compare size per page with a normal document from the same workflow.
- Zoom into text and try selecting it. A page that behaves as one picture points toward scan-as-image content.
- Look for full-bleed photographs, screenshots, colored paper, transparent artwork, maps, or dense diagrams.
- Open document properties in a trusted PDF inspector to review fonts and, where available, the space used by images and objects.
- Identify the document's role before editing: ordinary sharing copy, signed record, archival master, or print-production file.
Adobe's optimizer documentation separates images, fonts, transparency, removable objects, user data, and cleanup because each category creates different savings and different failure modes. That is a useful diagnostic model even when another tool exposes a smaller set of controls.1
Match seven causes to their signals, fixes, and risks
| Cause | Signal | Fix | Risk to check |
|---|---|---|---|
| High-resolution images | A few photos or screenshots stay sharp far beyond the intended viewing size. | Use PDFStay Image optimization for a screen copy, or export correctly sized images from the source application. | Small labels, crop detail, and print sharpness can soften.1 |
| Scan-as-image pages | Text will not select and each page contains paper texture, skew, or a camera shadow. | Use PDFStay Scanned PDF compression, starting with Clarity first or Balanced. | Searchable text, handwriting, codes, and thin marks may be lost or softened. |
| Inefficient color mode | Nominally black-and-white pages are stored as full color, including colored paper noise. | For a disposable scan copy, turn off Keep scans in color only if color carries no meaning. | Stamps, highlights, annotations, and color-coded evidence can become ambiguous.2 |
| Embedded fonts | A short, mostly textual PDF includes many typefaces or entire font programs. | Re-export with appropriate font subsetting in the source application; PDFStay has no font-unembedding control. | Removing fonts can substitute glyphs, change line breaks, or break multilingual text.1 |
| Redundant/hidden objects | The file has editing history, unused thumbnails, attachments, metadata, comments, or hidden application data. | Try Smart compression on a copy; use a specialist optimizer only when you understand each removal option. | Forms, tags, links, attachments, accessibility, or application workflows can break.1 |
| Layered or complex artwork | The PDF contains transparency, dense vectors, CAD paths, multiple layers, or alternate images. | Create a purpose-built export or flatten an approved copy in the source application; PDFStay has no layer-flattening switch. | Flattening can change transparency, spot colors, editability, line quality, and print behavior.1 |
| Already-compressed content | The PDF has already been optimized or its JPEG images are near their intended dimensions, yet the file remains over the target. | Keep the current file, split it, or change the delivery method; no PDFStay setting can guarantee another useful reduction. | A new lossy pass may add artifacts for very few saved bytes.1 |
See how content type changed three fixed results
PDFStay's fixed fixtures do not behave alike: the text-heavy record fell 48.33%, the scan-style record 91.26%, and the image-heavy record 81.74%. The dataset records input/output page counts of 4/4, 2/2, and 2/2 respectively. The methodology and JSON explicitly limit these numbers to known fixtures in the stated browser environment; they are diagnostic examples, not forecasts.3, 4
| Fixture | Recorded mode | Input bytes | Output bytes | Reduction | Input pages | Output pages | Readable-PDF check |
|---|---|---|---|---|---|---|---|
| text-heavy-pdf-v1 | standard | 11,762 bytes | 6,077 bytes | 48.33% | 4 | 4 | true |
| scan-style-pdf-v1 | scanned | 2,141,091 bytes | 187,107 bytes | 91.26% | 2 | 2 | true |
| image-heavy-pdf-v1 | image | 1,845,799 bytes | 337,064 bytes | 81.74% | 2 | 2 | true |
Fixed PDFStay fixture; not a promised compression rate or visual-quality score
Match the diagnosis to controls PDFStay actually exposes
| Route | Available control | Use it when | Verify |
|---|---|---|---|
| Compress PDF | Smart compression | The PDF is digital or mixed and should preserve selectable text where possible. | Text selection, links, forms, object behavior, and page count |
| Compress Images in PDF | Image optimization | Large embedded photos or screenshots dominate the file. | Image detail, captions, color transitions, and intended zoom or print size |
| Compress Scanned PDF | Scanned PDF compression | Pages are image-only scans and the output is an access or upload copy. | Fine text, handwriting, barcodes, stamps, searchability, and page order |
| Compress Scanned PDF advanced settings | Clarity first, Balanced, Smaller file, and Keep scans in color | You need to tune scan strength or retain meaningful color. | Every detail affected by the stronger setting or color conversion |
Require verification for signed, archival, and print-production PDFs
| Role | Keep | Verification |
|---|---|---|
| Signed PDF | The byte-identical signed original | Open the copy in the required signature validator and confirm every signature status and form requirement. |
| Archival PDF or preservation image master | The approved master, metadata, and capture record | Current institutional format, resolution, color, compression, metadata, and quality-control requirements2 |
| Print-production PDF | The supplied production file and output intent | Printer preflight, embedded fonts, image resolution, layers, transparency, color handling, and a proof1 |
FAQ
Can file size alone tell me why a PDF is large?
No. Size per page is a clue, but you still need to distinguish selectable content from scans and inspect images, fonts, objects, and the export workflow.
Will PDFStay Smart compression flatten my PDF layers?
Do not rely on it for that. PDFStay does not expose a layer-flattening control. Use the source application or a specialist production workflow and verify a separate copy.
When should I refuse to replace the original with a smaller PDF?
Keep the original whenever it is signed, archival, identity-related, legally significant, or intended for print production. A compressed copy can be useful for access, but it needs role-specific verification.
Sources
- Adobe Acrobat — PDF Optimizer settingsofficial technical documentationSource 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: