Released by the Joint Photographic Experts Group in 1992, JPEG solved a problem that mattered intensely at the time: photographs are enormous. Its answer was the Discrete Cosine Transform (DCT), which breaks an image into blocks of pixels and discards frequency information human vision struggles to detect anyway. A 10 MB camera shot becomes 500 KB with no visible damage at a sensible quality. That trade-off was revolutionary, and it built a thirty-year installed base.
01Why JPEG stuck
Every camera, phone, operating system, browser and CMS in existence reads JPEG. That universality is itself a feature — when compatibility is critical and file size is not, it is still a perfectly reasonable choice for photographs.
02What JPEG can’t do
The limits are real. JPEG has no alpha channel, so transparency is off the table — use PNG, WebP or AVIF for anything with a see-through background. It degrades with every re-save. Its artefacts — blocky ringing around hard edges, banding in gradients, muddy high-frequency detail — appear as quality drops below roughly 70 and become offensive below 50. It also has no animation and encodes 8 bits per channel.
03When to use it in 2026
For photographs going to the web, WebP and AVIF both produce smaller files at equivalent quality. Serve those via <picture> or srcset, with JPEG as the fallback for any legacy edge case. JPEG still shines as a universal interchange format for sending photos to clients, and as a safe export when compatibility beats everything else.
Strip EXIF metadata before publishing — location and device data travel with the file silently unless you remove them.