Image Compression Benchmarks: Real File Sizes Across Formats and Quality

Most image format comparisons you find online use a single test image and call it a day. We wanted real data — so we ran 50 different photographs, screenshots, and illustrations through our compressor and recorded every file size. Here are the results.

All tests were performed using our browser-based Image Compressor and format converters, which use the HTML5 Canvas API with standard encoding libraries. No server processing — pure browser rendering.

Test Methodology

We used 50 images across three categories:

• 20 photographs — Landscapes, portraits, street photography (2–8MP, JPG format)
• 15 screenshots — Web pages, app interfaces, code editors (1–3MP, PNG format)
• 15 illustrations — Logos, icons, digital art, diagrams (0.5–2MP, PNG format)

Each image was compressed at quality settings of 20%, 40%, 60%, 80%, and 100% (lossless for applicable formats). File sizes were measured before and after compression.

Benchmark 1: Format Comparison at Quality 85%

We took a representative 1920×1080 photograph (5.2 MB original) and converted it to every format at quality 85%:

Key finding: AVIF produces the smallest files — 46% smaller than JPG and 22% smaller than WebP at the same quality level. PNG is the largest by far, making it a poor choice for web delivery of photographs.

Benchmark 2: Quality Settings Impact (JPG)

How much does lowering quality actually save? We compressed the same 1920×1080 photograph at every quality level:

The sweet spot is quality 80–85%. You get 95-96% file size reduction with zero visible quality loss at web viewing sizes. Dropping to 60% saves another 36% of space but starts showing artifacts in gradients and fine details.

Benchmark 3: Image Type Matters

Not all images compress the same way. We tested three different image types at the same quality setting (80%) to show the difference:

Photographs compress the most because they contain more "information" that lossy algorithms can discard. Screenshots and illustrations have sharper edges and fewer gradients, so they compress less aggressively but still see dramatic savings.

Benchmark 4: Batch Compression to Target Size

We ran 10 mixed-format images through our Compress to 100KB tool to test how the target-size algorithm performs:

The target-size algorithm consistently hits the 100KB target. For images that are already small, it preserves more quality. For large images, it applies more aggressive compression to hit the target.

Benchmark 5: Browser Speed

Since our tools run entirely in the browser, processing speed depends on your device. We measured compression times on three representative devices:

Even on a budget Android phone, compressing 10 photos takes under 3 seconds. Server-based tools typically take 5–15 seconds for the same operation due to upload/download latency.

What These Numbers Mean for You

Here are the practical takeaways from our benchmarks:

For Web Performance

• Use WebP for photos — 97% smaller files than the original, 30% smaller than JPG at the same quality
• Use quality 80–85% — The sweet spot where file size drops dramatically but quality stays perfect at web sizes
• PNG only for transparency — PNG files are 3× larger than WebP for photographs. Use PNG only when you need lossless transparency

For Government Forms and Portals

• 100KB is achievable for almost any photo — Even 5MB photos compress to under 100KB at quality 60–70%
• Batch compression saves time — 10 photos in under 3 seconds, even on mobile
• Target-size tools are reliable — Our 100KB compressor hit the target 100% of the time in testing

For Privacy

• All this happens in your browser — Zero upload time, zero server processing, zero data exposure
• Speed is comparable to desktop apps — Browser-based compression at 0.12s per photo rivals desktop software

Ready to test it yourself? Try our Image Compressor or Compress to 100KB tool and see the numbers firsthand.