Estimate PDF compression before you compress
Compressing a PDF blindly is a gamble: sometimes the file drops from 30 MB to 2 MB, sometimes it barely changes and the text turns blurry. The PDF Compression Analyzer tells you in advance how much each optimisation technique could save on your file, so you can decide what is worth doing.
It finds where every image is placed on the page and calculates its effective resolution, then actually recompresses each image in your browser at the resolution and JPEG quality you choose to measure the result. It also detects images and fonts stored more than once, streams saved with no compression at all, fonts embedded in full instead of as subsets, missing object streams, page thumbnails and leftover objects.
All of this runs locally. The PDF is never uploaded, and nothing in it is changed — this is a report, not a converter.
How to use the PDF Compression Analyzer
- Choose a target image resolution: 150 DPI suits on-screen reading and office printing, 300 DPI professional print.
- Set the JPEG quality used for the test (75% is a sensible default).
- Drop your PDF onto the upload area. The analysis starts automatically and shows its progress.
- Read the “What you’d save with each technique” table and the estimated final size.
- Check the image and font tables to see which objects matter, change the settings to compare, and download the report as CSV if you need it.
PDF Compression Analyzer features
Real test compression
The largest images (up to 30) are decoded and re-encoded with your browser’s JPEG encoder at the target size, so estimates are measured rather than guessed.
Effective DPI per image
A content-stream interpreter follows each page’s transformations to find how big every image is drawn, and therefore how many pixels it really needs.
Duplicate detection
Images and font programs are hashed with SHA-256; identical copies stored separately are listed with the space they waste.
Lossless opportunities
Uncompressed streams are test-compressed with Deflate, and missing object streams are measured, so you see savings that cost no quality.
Font embedding check
Every embedded font is marked as subset or full, with a rough estimate of what subsetting would save.
Compared with Compress PDF
A few pages are rendered the way Compress PDF does it, so you can compare page rasterisation with targeted optimisation.
How the estimates are calculated
Effective resolution. An image of 1,600 × 1,200 pixels drawn 2 inches wide on the page has 1,600 ÷ 2 = 800 pixels per inch. At a 150 DPI target only 300 × 225 pixels are needed, so the tool scales a copy to that size, encodes it as JPEG at your chosen quality and records the real byte count. Images that are already at or below the target are tested at their current size and kept if recompression would not help.
Extrapolation. To keep the analysis quick on documents with hundreds of pictures, only the 30 largest images are test-compressed; the rest use the median bytes-per-pixel of the tested ones, and are labelled “estimated from similar images”.
Fonts. A simple PDF font can address at most 256 characters, so a fully embedded TrueType font with thousands of glyphs carries mostly unused data. The estimate scales the font size by the share of glyphs that can be used — a rough guide, since tables such as kerning and hinting do not shrink proportionally.
Which techniques lose quality?
| Technique | Quality impact | Typical gain |
|---|---|---|
| Downsampling + JPEG recompression | Lossy — fine detail and sharp edges soften | Large on scans and photos |
| Storing duplicates once | None | Large on merged documents |
| Compressing uncompressed streams | None | Moderate, depends on the generator |
| Font subsetting | None for viewing; editing text later may lack glyphs | Moderate on documents with CJK or large fonts |
| Object streams | None (needs PDF 1.5+ readers) | Small to moderate |
| Removing thumbnails and unused objects | None | Small, occasionally large |
Turning the report into a smaller file
If images dominate and the document is a scan or photo collection, Compress PDF gives you a small file in one step — remember that it renders each page as an image, so selectable text, links and form fields do not survive. For digital documents where text must stay selectable, re-export from the original application with image downsampling and font subsetting turned on, using the DPI the report shows is sufficient.
To check print readiness rather than file size, the PDF Resolution Checker flags images below 150 or 300 PPI, and the PDF Size Analyzer gives a full breakdown of what the file contains.
PDF Compression Analyzer FAQ
How much can I compress my PDF?
It depends on what is inside. Drop the file into the PDF Compression Analyzer: the total at the top combines all techniques, and the table shows each technique’s share. Scanned documents often shrink by 70–95%, while text-only reports may only lose a few percent.
Are the estimates exact?
Image savings for the largest images are measured by really recompressing them in your browser, but a PDF optimiser may use a different JPEG encoder, so expect results within a reasonable margin. Font subsetting figures are rough estimates.
What DPI should I choose?
150 DPI is plenty for reading on screen and printing on office printers. Use 300 DPI for professional print, and 72–96 DPI only for files meant purely for screens or email.
Why does an image show “display size unknown”?
The image is used somewhere the analyser does not follow — for example inside a pattern or an annotation appearance. It is then assumed to need its full resolution, so the estimate stays conservative.
Does this tool change my PDF?
No. It only reads the file and reports potential savings. Use Compress PDF or your PDF editor’s optimise feature to create a smaller copy.
Is my file uploaded?
No. Parsing, decoding and test compression all happen in your browser, so the document stays on your device.