How to Compress a PDF Without Losing Quality
How to Compress a PDF Without Losing Quality
The phrase "compress without losing quality" requires a precise definition. In PDF compression, text and vector graphics never degrade — they are mathematically lossless by nature. Embedded raster images (photographs, scanned pages) are the only content that can degrade under lossy compression. Understanding this distinction lets you compress aggressively where it is safe and conservatively where it matters.
The Four PDF Compression Techniques
A PDF compression engine typically applies some combination of these four techniques:
1. Image Recompression (Lossy)
This is the primary lever. Every embedded JPEG or rasterised image is re-encoded at a lower quality setting. At quality 85, the visual difference from quality 100 is imperceptible for most photographic content. At quality 70, edges may show slight ringing artefacts. Below quality 60, degradation becomes visible on screen.
Impact: 40–75% size reduction for image-heavy documents. Risk: Visible on close inspection of photographs at high compression levels.
2. Image Downsampling
Images embedded at 300 DPI are downsampled to 150 DPI or 96 DPI. Pixels are discarded permanently. For a document that will only ever be viewed on screen, this is almost always safe. For a document that must be printed at high quality, it must be avoided.
Impact: 50–80% size reduction for high-resolution scanned documents. Risk: Print quality is reduced.
3. Flattening Revision History (Lossless)
Each incremental save of a PDF appends a new revision layer to the file. The old objects remain. Flattening writes a single clean copy of the current state, discarding all revision history. No visual content is changed.
Impact: 10–50% size reduction on repeatedly-edited documents. Risk: None — revision history is not rendered content.
4. Metadata and Font Subsetting (Lossless)
Removing XMP metadata, colour profiles, preview thumbnails, and unused font glyphs. Font subsetting keeps only the characters actually used in the document rather than the full character set.
Impact: 5–20% size reduction. Risk: None for viewing. Subsetting fonts can occasionally cause issues if a downstream PDF editor needs to add new characters in the same typeface.
⚡ Fix This Instantly: Our Compress PDF tool applies all four techniques with one click. Choose Low to preserve maximum quality, Medium for the best size/quality balance, or High for the smallest possible file.
Compression Level Reference
| Level | Techniques Applied | Quality Impact | Typical Reduction |
|---|---|---|---|
| Low | Metadata removal, font subsetting, mild recompression | Imperceptible | 10–30% |
| Medium | All above + moderate image recompression (Q80) | Minimal | 30–60% |
| High | All above + aggressive recompression (Q60) + downsampling | Noticeable on images | 60–80% |
For most use cases — emailing contracts, submitting scanned forms — Medium is the correct choice.
Advertisement
What Cannot Be Compressed
- Text streams — already stored as compressed objects in any modern PDF. Further compression yields no gain.
- Vector graphics — mathematical descriptions of curves and fills, inherently compact.
- Linearised PDFs optimised for web — already structured for minimal load; re-compression gains are small.
- Previously compressed PDFs — if a PDF was already compressed by another tool, you are compressing a compressed output. Gains will be 5–15% at best.
Verifying Quality After Compression
After compression, open the PDF and:
- Zoom in to 150% on text — it should still be sharp with no pixelation (text is never degraded)
- Zoom in to 100–150% on any embedded photographs — assess whether visible artefacts are acceptable for your use case
- Print one page if the document will be submitted in printed form
If quality is insufficient, use a lower compression level and re-compress from the original source file — never re-compress an already-compressed PDF.