What's Hiding in Your Photos: EXIF Metadata, and Why Compression Isn't Always Lossless

Two things about a photo file are easy to overlook until they matter: what's embedded in it beyond the visible pixels, and what actually happens to the pixel data when you "compress" it. Both are more specific than most people assume.

What EXIF metadata actually contains

Every photo from a modern phone or camera carries EXIF (Exchangeable Image File Format) data — a block of metadata embedded alongside the pixel data. This can include GPS coordinates (if location services were on), the exact camera or phone model, the lens used, exposure settings, and the precise timestamp the photo was taken. None of this is visible when you look at the image; it's only visible to software that reads the file's metadata specifically.

EXIF Editor reads this with exifr, a real JavaScript EXIF parser, and lets you inspect what's actually embedded before you decide whether to share the file as-is. Stripping it works by redrawing the image onto a canvas and re-exporting it — the rebuild is what discards the metadata, since a freshly drawn canvas has no memory of the original file's embedded data. Worth being specific about GPS: it's shown as raw latitude/longitude coordinates, not a clickable map link — you'd copy those numbers into a map yourself if you wanted to see the location visually.

Why "compress" doesn't mean the same thing for every format

This is where a lot of confusion comes from, because two formats handle recompression completely differently:

• PNG is a lossless format — reprocessing a PNG through Image Compressor or the metadata-stripping rebuild in EXIF Editor doesn't change a single pixel value, only the file's structural encoding.
• JPEG and WebP are lossy by nature — any tool that recompresses them, ours included, re-encodes the image data at a real compression setting (fixed at roughly 80% quality internally), which is not bit-for-bit identical to the source even though the visual difference is typically minimal at normal viewing sizes.

This means "lossless" is only accurate for PNG specifically. A JPEG that's been stripped of metadata or compressed has genuinely been re-encoded — usually with no visible quality loss, but not literally unchanged data.

Compressing by size or resolution, not a quality percentage

Image Compressor doesn't expose a quality-percentage slider — it works by Target Max Size (a size in MB you're aiming for) and Target Resolution (fixed options: 800, 1080, 1920, or 2560px on the longest side), using browser-image-compression running in a Web Worker so the compression itself doesn't freeze the browser tab while it processes. It also handles one image at a time — there's no batch/multi-file mode today.

Two more places pixel-level work shows up

Favicon Generator redraws your source image at each size a browser or device expects (16, 32, 180, 192, and 512px) using center-crop if the source isn't already square — so a rectangular logo gets cropped to a square region rather than stretched into a distorted one. Image Color Palette samples an image and returns exactly 6 dominant colors as hex codes via local pixel sampling — a fixed count, not a configurable 5/8/12 option, and hex only (no RGB or CSS-variable export today).

The Bottom Line

A photo carries more than its visible pixels, and "compress" or "strip metadata" means something specifically different depending on the format you're working with. Knowing that PNG recompression is genuinely lossless while JPEG/WebP recompression is a real (if usually invisible) trade-off — and that metadata stripping works by rebuilding the file rather than editing fields in place — sets the right expectations before you rely on either for something that matters, like scrubbing location data before posting a photo publicly.