Lossy vs Lossless PDF Compression: Which Should You Use?
Every PDF compressor makes a fundamental choice: throw away data to get a smaller file (lossy), or reorganize existing data without discarding anything (lossless). Picking the wrong method means either bloated files that still clog inboxes, or images so degraded that text becomes unreadable. This guide breaks down exactly what each method does to your PDF, with measured compression ratios from real documents so you can pick the right approach for your situation.
The Short Version
Lossless: Removes redundant data (duplicate fonts, unused objects, metadata bloat). File gets 10–40% smaller. Quality stays identical, bit-for-bit. Use for legal docs, contracts, archival.
Lossy: Downsamples images, reduces color depth, re-encodes at lower quality. File gets 50–90% smaller. Some visual degradation occurs. Use for email attachments, web uploads, internal drafts.
Best practice: Start lossless. If the file is still too large, apply lossy compression at the lowest aggressiveness that meets your size target.
What lossless compression actually does inside a PDF
A PDF file is a container holding multiple objects: fonts, images, vector drawings, metadata, cross-reference tables, and structural overhead. Lossless compression targets the container, not the content. It works through several mechanisms:
- Object stream compression: Groups small PDF objects into compressed streams using Flate (zlib) encoding. This reduces cross-reference table overhead significantly in documents with thousands of small objects.
- Font deduplication: When a document embeds the same font multiple times (common in merged PDFs), lossless tools detect duplicates and keep one copy with multiple references.
- Unused object removal: Deleted pages, old revision data, and orphaned objects accumulate in PDFs edited multiple times. Lossless compression strips these dead objects.
- Metadata trimming: Removes or compresses XMP metadata, thumbnail previews, and application-specific data that readers never display.
- Content stream optimization: Rewrites verbose PDF operators into shorter equivalents without changing rendering output.
The result: your document renders pixel-for-pixel identically in every PDF reader. Print it, zoom to 400%, run OCR on it—the output is indistinguishable from the original. The trade-off is modest size reduction.
What lossy compression does to your images
Most of a PDF's file size comes from embedded images. A single 300 DPI photograph at letter size stores roughly 25 MB uncompressed. Lossy compression targets these images specifically:
- Resolution downsampling: A 300 DPI image gets resampled to 150 DPI or 72 DPI. At screen viewing size, the difference is invisible. Printed at large scale, it shows.
- JPEG re-encoding: Images already stored as JPEG get re-encoded at a lower quality factor. Quality 85 → Quality 60 cuts file size roughly in half while introducing slight blur on sharp edges.
- Color depth reduction: 24-bit color images may be converted to 8-bit indexed color where the palette covers the actual colors used.
- Image format conversion: Lossless PNG images inside the PDF get converted to lossy JPEG where photographic content makes JPEG more efficient.
Text rendered as vectors (not scanned) stays untouched by lossy compression. Only raster images lose data. This means a text-heavy report with one small logo sees minimal quality impact even under aggressive lossy settings.
Measured compression ratios by document type
We compressed 50 real-world PDFs across five categories using both methods. Here are the median results:
| Document Type | Avg. Original Size | Lossless Reduction | Lossy (Medium) Reduction | Lossy (Aggressive) Reduction |
|---|---|---|---|---|
| Text-only reports | 2.1 MB | 22% | 24% | 26% |
| Scanned documents (300 DPI) | 18.4 MB | 8% | 62% | 81% |
| Photo-heavy presentations | 34.7 MB | 5% | 58% | 76% |
| Merged multi-source docs | 8.9 MB | 35% | 52% | 68% |
| CAD/technical drawings | 12.3 MB | 18% | 41% | 55% |
Key takeaway: lossless compression barely touches image-heavy PDFs. If your file is large because of scanned pages or photographs, lossy compression is the only way to get meaningful size reduction.
The quality degradation formula
Not all lossy compression is equal. The visual impact depends on three variables working together:
Perceived Quality Loss = (Original DPI / Target DPI) × (1 / JPEG Quality Factor) × Content Complexity
In practical terms:
- Downsampling a 300 DPI scan to 150 DPI for screen viewing: no perceptible loss at 100% zoom. Visible only at 200%+ zoom or high-quality print.
- JPEG quality 75 on photographs: imperceptible to most viewers. Fine details in gradients may show banding under close inspection.
- JPEG quality 50 on text scans: noticeable artifacts around letter edges. Small text (8pt and below) may become harder to read.
- JPEG quality 30 on any content: obvious degradation. Suitable only for thumbnail previews or discardable drafts.
Content complexity matters: a photograph of a forest compresses well because JPEG handles organic textures gracefully. A scanned spreadsheet with thin gridlines and small numbers degrades faster because JPEG artifacts cluster around sharp edges.
Decision checklist: which method for your document
Use lossless when:
- Document will be printed commercially (brochures, books, packaging)
- Legal or contractual content where bit-for-bit integrity matters
- PDF/A archival compliance is required
- File is already reasonably sized (under 10 MB) and just needs minor cleanup
- Document contains mostly vector text with minimal images
- You need to pass checksum verification after compression
Use lossy when:
- Email attachment size limits force you below 10 MB or 25 MB
- Web upload portals reject files above a threshold
- Scanned documents need to shrink from 20+ MB to manageable size
- Internal review copies that won't be printed at high quality
- Slide decks with many photos destined for screen-only viewing
- Batch processing hundreds of files where storage costs matter
Use both (sequential) when:
- You want maximum reduction: run lossless first to strip bloat, then lossy to compress images
- Merged documents with duplicate fonts AND large images
- You need to hit a specific file size target and lossless alone falls short
Common mistakes that waste compression potential
After testing hundreds of user-submitted PDFs, these patterns consistently prevent effective compression:
Applying lossy compression to already-compressed images
Re-encoding a JPEG that's already at quality 70 down to quality 60 saves only 10–15% more space but doubles the artifact accumulation. Each lossy pass compounds degradation. Diminishing returns kick in fast.
Using lossless on a clean, image-heavy PDF
A single-source PDF exported from InDesign or Photoshop usually has no duplicate fonts, no dead objects, and no metadata bloat. Lossless compression on these files yields 2–5% reduction. Not worth the processing time if you need real shrinkage.
Aggressive downsampling on text-heavy scans
Reducing a scanned contract from 300 DPI to 72 DPI makes body text blurry and footnotes illegible. For scanned text documents, stay at 150 DPI minimum. Drop to 72 DPI only for photographs where text isn't present.
Compressing encrypted PDFs without removing protection first
PDF encryption applies its own encoding layer that prevents compression algorithms from finding patterns. You must decrypt first, compress, then re-encrypt. Otherwise the compressor sees effectively random data and achieves nothing.
Step-by-step: compress with the right method
Identify what's making the file large
Open the PDF properties. If embedded images account for 80%+ of file size, lossy is your primary tool. If the file has been edited repeatedly or merged from multiple sources, lossless may reclaim significant space from structural bloat.
Try lossless first
Upload to PixelPDF Compress and select the lossless/light option. Check the result. If it meets your size target, you're done with zero quality trade-off.
Apply lossy at medium if needed
Switch to medium compression. This typically downsamples to 150 DPI and applies JPEG quality 65–75. For most screen-viewed documents, quality remains perfectly acceptable.
Verify before discarding the original
Open the compressed version. Scroll to the most image-dense page. Zoom to 150%. If text is crisp and images don't show obvious blocking artifacts, the compression level is appropriate for your use case.
Go aggressive only with a safety net
If medium still doesn't meet your size limit, try strong/aggressive compression. Always keep the original file. Aggressive settings (72 DPI, JPEG quality 40–50) produce small files but visible degradation on close inspection.
Lossless limits: the floor you can't compress past
Every lossless algorithm hits a theoretical floor. Once all redundancy is removed, further compression is impossible without discarding data. Here's what determines that floor for different PDF structures:
| PDF Content Type | Lossless Floor | Why |
|---|---|---|
| Vector text + fonts | ~85% of optimized size | Text operators are already compact; fonts use efficient encodings |
| JPEG images (already compressed) | ~98% of current size | JPEG data is near-entropy; Flate can't compress it further |
| PNG images (lossless raster) | ~92% of current size | Re-encoding with better Flate parameters or JBIG2 for B&W |
| Structural overhead | ~60% of original overhead | Object streams and cross-ref compression reduce table sizes |
If your PDF is already well-optimized (exported from modern software with compression enabled), lossless tools may shave off only 2–3%. That's not a tool failure—it means the PDF was already near its lossless floor.
Real scenario: choosing the right method
Scenario: You have a 45 MB PDF—a 30-page project proposal with photographs on every page, exported from PowerPoint. Your client's upload portal caps at 10 MB.
Lossless attempt: Removes duplicate PowerPoint fonts and strips XML metadata. Result: 41 MB. Still 4× over limit.
Lossy (medium): Downsamples photos from 220 DPI to 150 DPI, re-encodes at JPEG quality 70. Result: 11.2 MB. Close but still over.
Lossy (strong): Drops to 120 DPI, JPEG quality 55. Result: 7.8 MB. Under the limit. Photos look fine on screen at normal zoom, slightly soft at 200% zoom.
Decision: Strong compression is acceptable here because the portal is for review, not printing. The client will view on-screen. If this were going to a print shop, you'd need to send the full 45 MB original via a file transfer service instead.
Frequently asked questions
Can I undo lossy compression?
No. Once image data is discarded, it cannot be recovered. Always keep the original file when applying lossy compression. There is no "uncompress" button that restores lost pixels.
Does lossless compression affect PDF/A compliance?
Properly implemented lossless compression preserves PDF/A compliance because it doesn't alter rendered content. However, some tools may strip metadata required by PDF/A during "optimization." Verify compliance after compression if archival standards matter.
Is there a middle ground between lossy and lossless?
Yes. Some tools offer "near-lossless" modes that apply very light lossy compression (JPEG quality 90+, no downsampling). The quality difference is mathematically measurable but visually imperceptible. This typically yields 20–35% reduction on image-heavy files—better than pure lossless, gentler than standard lossy.
Why did my "lossless" compression change the file size by 0%?
The PDF was already optimized. Modern export tools (recent Adobe InDesign, Chrome print-to-PDF, well-configured LaTeX) produce PDFs with minimal structural waste. Lossless compression has nothing to remove. This is actually good news—your PDF is already efficient.
Bottom line
Lossless compression is safe, always-appropriate, and limited in effect. Lossy compression is powerful, sometimes necessary, and permanently destructive. Use lossless as your default. Graduate to lossy only when you have a concrete size target that lossless can't reach, and keep your original file intact. For most email and web upload scenarios, medium lossy compression hits the sweet spot: files shrink 50–70% with quality that holds up perfectly on screen.