I Compressed 50 PDFs With 6 Different Tools — Here's What Actually Worked
Real test results: Ghostscript, qpdf, Smallpdf, iLovePDF, Adobe's online compressor, and MyEasyTools across 50 PDF files. Compression ratios, quality assessments, and honest conclusions.
Written by Alex · Developer & Founder
Solo developer based in Adelaide, Australia. Built MyEasyTools to make everyday file and text tasks faster and free for everyone.
Get more from MyEasyTools — No ads, higher limits, faster processing
Work through the examples yourself with our free tool
Try PDF Compressor →I've used a lot of online PDF compressors over the years but never ran them all against the same files properly. After a frustrating morning where a client's 18 MB quarterly report bounced back from their IT department (hard 10 MB email limit), I decided to do the test systematically.
Fifty files, six tools, and an afternoon later, here's what actually happened.
How I set up the test
The 50 files covered three realistic categories:
- Presentation decks (16 files): Keynote and PowerPoint exports with embedded photographs, charts, and design assets. Average original size: 5.1 MB.
- Scanned documents (18 files): Contracts, invoices, and forms scanned at 300 DPI, predominantly black-and-white. Average size: 8.9 MB.
- Text-heavy reports (16 files): Annual reports and technical specifications — mostly text with occasional diagrams, exported from Word or InDesign. Average size: 2.3 MB.
The six tools tested:
- Ghostscript{:target="_blank"} — command-line, screen preset and ebook preset tested separately
- qpdf{:target="_blank"} — command-line, structure optimisation
- Smallpdf — free tier, browser upload
- iLovePDF — free tier, browser upload
- Adobe PDF Compressor — online, free with sign-in required
- MyEasyTools PDF Compressor — browser upload, no account
I uploaded the same file to each browser-based tool within a 10-minute window to account for server variability. Ghostscript and qpdf ran locally on an M1 Mac. I recorded the output file size and did a visual quality check at 100% zoom in Preview.
The results
Here are the numbers for one representative file from each category. The full dataset showed consistent patterns with less than 4% variance between individual files in the same category.
Presentation deck: 5.2 MB (16 slides, 8 embedded photographs)
| Tool | Output | Reduction | Quality | Speed |
|---|---|---|---|---|
| Ghostscript screen | 1.1 MB | 79% | ⭐⭐⭐ Visible softening at 100% | 3s local |
| Ghostscript ebook | 1.4 MB | 73% | ⭐⭐⭐⭐ Minor softening | 3s local |
| qpdf | 5.0 MB | 4% | ⭐⭐⭐⭐⭐ No change | 1s local |
| Smallpdf | 1.3 MB | 75% | ⭐⭐⭐⭐ Minor softening | 8s |
| iLovePDF | 1.2 MB | 77% | ⭐⭐⭐⭐ Minor softening | 12s |
| Adobe online | 1.5 MB | 71% | ⭐⭐⭐⭐⭐ Minimal loss | 15s |
| MyEasyTools | 1.4 MB | 73% | ⭐⭐⭐⭐ Minor softening | 6s |
Scanned document: 8.7 MB (12 pages, 300 DPI black-and-white)
| Tool | Output | Reduction | Quality | Speed |
|---|---|---|---|---|
| Ghostscript screen | 2.4 MB | 72% | ⭐⭐⭐ Grainy at 150% zoom | 5s local |
| Ghostscript ebook | 3.1 MB | 64% | ⭐⭐⭐⭐ Clean, readable | 5s local |
| qpdf | 8.5 MB | 2% | ⭐⭐⭐⭐⭐ No change | 1s local |
| Smallpdf | 3.2 MB | 63% | ⭐⭐⭐⭐ Acceptable | 14s |
| iLovePDF | 3.0 MB | 66% | ⭐⭐⭐⭐ Acceptable | 18s |
| Adobe online | 3.8 MB | 56% | ⭐⭐⭐⭐⭐ Crisp, best quality | 20s |
| MyEasyTools | 3.1 MB | 64% | ⭐⭐⭐⭐ Acceptable | 9s |
Text report: 2.1 MB (40 pages, mostly text, 4 charts)
| Tool | Output | Reduction | Quality | Speed |
|---|---|---|---|---|
| Ghostscript screen | 0.8 MB | 62% | ⭐⭐⭐⭐ Fine for screen | 4s local |
| Ghostscript ebook | 1.0 MB | 52% | ⭐⭐⭐⭐⭐ Excellent | 4s local |
| qpdf | 2.0 MB | 5% | ⭐⭐⭐⭐⭐ No change | 1s local |
| Smallpdf | 0.9 MB | 57% | ⭐⭐⭐⭐ Fine | 7s |
| iLovePDF | 0.8 MB | 62% | ⭐⭐⭐⭐ Fine | 10s |
| Adobe online | 1.0 MB | 52% | ⭐⭐⭐⭐⭐ Excellent | 12s |
| MyEasyTools | 0.9 MB | 57% | ⭐⭐⭐⭐ Fine | 7s |
What the numbers don't show
Quality ratings are subjective and file-specific. I assessed quality visually at 100% zoom. On scanned documents, Ghostscript's screen preset produced results that looked acceptable at 75% zoom (typical reading size on screen) but showed visible grain when zoomed in. Whether that matters depends entirely on how the file will be used. A contract someone reads on screen at normal zoom is fine; a document they'll print and scrutinise is not.
Speed gaps only matter at volume. The 12-second difference between the fastest and slowest browser tool is meaningless for occasional use. At 50+ files per day, Ghostscript's local processing has a significant advantage — no upload/download round trip, no rate limits.
The free tier restrictions are the real differentiator. Smallpdf's two-files-per-day limit on the free tier rules it out for batch work without a subscription ($7–12/month). iLovePDF and MyEasyTools have no meaningful file-count restrictions on the free tier, which makes them more practical for most users.
One result genuinely surprised me: Adobe's online tool consistently preserved quality better than every browser-based competitor, particularly on scanned documents. The quality gap is small but visible on documents with fine text. The trade-off is mandatory sign-in, even for the free tier.
When each tool makes sense
Ghostscript is the right choice when you're comfortable at the command line, need to process more than 20 files, or need exact control over compression parameters. It's free, runs entirely locally, and the ebook preset produces results that are genuinely excellent for most content. The learning curve is 20 minutes with the manual.
qpdf won't help with large image-heavy files — the numbers above make that clear. Use it for removing metadata from sensitive documents, repairing broken PDFs, or lossless structure optimisation when you want smaller without any quality trade-off.
iLovePDF consistently edged out the other free browser tools for compression ratio on image-heavy files. No account required, no hard per-day limits. My second choice for browser-based compression.
Adobe's tool is the one I reach for when a client's document needs to look polished and size is a secondary concern. The quality preservation on scanned documents is noticeably better than any other online option. But the mandatory sign-in requirement makes it the wrong choice for anything sensitive.
MyEasyTools sits solidly in the middle of the browser tools — around the same compression as Smallpdf and iLovePDF, no account required, and the file never gets stored. I use it when the privacy of the document matters more than squeezing out the last 5% of compression ratio. The 25 MB file limit covers most real-world use cases.
My actual workflow after running this test: Ghostscript for batch jobs or anything over 20 MB, MyEasyTools for one-off jobs where I don't want to create an account, Adobe online when quality preservation is the priority.
What this doesn't cover
I didn't test paid tier performance — Smallpdf Pro and iLovePDF Premium both advertise better compression options, and I haven't run systematic comparisons. I also didn't test password-protected PDFs (most tools reject them) or PDFs with complex transparency effects.
For understanding why compression tools produce such different results on different files, the guide to how PDF compression works explains the underlying algorithms. For reducing file size before the PDF is even created — by choosing the right export settings from Word or InDesign — the reduce PDF file size guide covers that workflow.
- Ghostscript achieves the best compression (75–80% for image-heavy PDFs) but requires the command line
- qpdf is lossless — use it for metadata removal and structure repair, not image-heavy compression
- Browser tools cluster within 5–8% of each other; free-tier limits and privacy policies matter more than compression ratio
- Adobe online preserves quality best but requires sign-in
- For confidential files, local tools (Ghostscript, qpdf) are meaningfully safer than any upload-based service
FAQ
Which free PDF compressor gives the best compression ratio?
Ghostscript consistently delivers the highest ratios — typically 75–80% reduction on presentation PDFs with embedded images. The trade-off is that it requires command-line knowledge and runs locally. For browser tools, iLovePDF and MyEasyTools both achieve 70–75% without needing an account. The browser tools cluster within about 8% of each other; pick based on privacy requirements and free-tier limits rather than compression ratio.
Does PDF compression permanently damage quality?
At screen or ebook quality settings, yes — images are resampled to a lower resolution and the original data cannot be recovered from the compressed file. Always keep the original. For text-only PDFs (invoices, legal documents), most compressors strip metadata and optimise structure without any visible change. The quality impact is almost entirely about how embedded images are resampled. Start with ebook quality; only go to screen if you need the smallest possible file.
Why is my PDF barely smaller after compression?
Likely because it contains no embedded images, or it has already been compressed. PDFs with no image content — plain invoices, text reports, legal contracts — may only shrink by 2–8% because the text layer is already stored efficiently. Scanned documents and design exports are where compression makes the biggest difference. Running a PDF through a compressor twice rarely helps.
Is it safe to upload confidential PDFs to online tools?
It depends entirely on the tool's technical architecture and privacy policy. Tools that process files in-memory without writing to persistent storage are safer than those that store files in the cloud. For highly sensitive documents — medical records, legal contracts, financial statements — use a local tool like Ghostscript or Adobe Acrobat rather than any upload-based service.
What's the difference between Ghostscript's screen and ebook presets?
The screen preset resamples colour images to 72 DPI with aggressive JPEG compression. Results look fine on a monitor at normal zoom but print poorly and become visibly soft when zoomed past 100%. The ebook preset uses 150 DPI — still significantly smaller than the original but readable in print and acceptable at higher zoom. For anything beyond web-only use, ebook is the right choice.