Image Tools Research
Original experiments on image compression and image privacy — measured, not rehashed.
Why This Section Exists
Most advice about image compression repeats the same three sentences: "quality 60–80 is fine", "WebP is smaller", "compress before uploading". Some of it is true, but almost none of it comes with numbers you can check.
Every study in this section follows the same rules:
- Real images, stated in advance. We publish exactly what our test set contains — how many images, what kind, what dimensions — so you can judge how well it matches your use case.
- The same codecs your browser uses. All encoding is done with the browser's native image encoders — the same engine behind our compression tool. When we report a file size, that's the file you would actually get.
- Raw numbers alongside conclusions. Every table shows the measurements, not just the takeaway. If you disagree with our interpretation, the data is there for you to re-read.
- Reproducible. Each study states its method precisely enough that a careful reader could repeat it and should get similar numbers.
Studies are added as they are completed — we publish one when it's done, not on a schedule.
Published Studies
JPEG Quality 50–90: What 20 Real Images Actually Told Us
We re-encoded 20 images — photos, UI screenshots, text documents, and flat illustrations — at quality 50 through 90 using the browser's own encoder, and measured both file size and structural similarity (SSIM). Includes one result most compression guides get wrong: when quality 70 produces larger files than quality 50 with worse fidelity, and why.
WebP vs JPEG: What the Format Switch Actually Saves
Same 20 images, same browser encoders. WebP beats JPEG on both size and fidelity for screenshots, documents and graphics — including one lossless result 3.6× smaller than PNG. For photos, the honest answer (at matched quality) is more nuanced than "WebP is smaller".
Does Resizing Before Compression Beat Compression Alone?
10 photos × 4 widths × 5 quality levels. Resizing to display size first cuts files up to 12× at equal visible quality — and file size turns out to track pixel count almost exactly. But the full-resolution route has a hidden quality advantage: browsers hide JPEG artifacts when they shrink images for display.
What EXIF Metadata Reveals About a Photo
26 metadata fields — including exact GPS coordinates — read from one phone photo. Then a measured survival matrix: which operations (browser encoding, re-saving, format conversion) actually strip the private fields. Six out of seven do, no special privacy mode required.
Browser-Side vs Server-Side Processing: What Leaves Your Computer?
We instrumented a real browser-side compression page and a server-side upload flow with the same files: 0 requests and 0 bytes uploaded versus a 4.5 MB round trip for a 3.6 MB photo. Includes a 30-second DevTools test to verify any tool yourself — including ours.
Use the Tools Behind the Studies
Every measurement in this section was produced with the site's own browser-based tools. Try them on your own images: