JPEG turned 32 in 2024. PNG has been around since 1996. Both feel permanent because they survived long enough to become infrastructure. But the history of digital imaging is actually littered with formats that had real technical merit, attracted real adoption, and then quietly stopped mattering. Understanding why they failed — or just faded — tells you something useful about how formats actually succeed, and what “open standard” or “better compression” has historically been worth when the ecosystem doesn’t follow.
Here are ten formats worth knowing about, even if you’ll never save a file in any of them.
The Contenders That Lost
PCD — Kodak Photo CD
Launched in 1992, Kodak’s Photo CD format was genuinely ahead of its schedule. It stored a single image at five different resolutions simultaneously — 192×128 up to 3072×2048 — inside a single .pcd file, using a custom hierarchical encoding Kodak called Image Pac. The color model was a proprietary YCC space (Luma plus two chroma channels), and the format supported 24-bit color before most consumer monitors could render it.
Kodak designed it for a world where you’d scan film at a photo lab, put the results on a disc, and view or print from that disc at home. The infrastructure arrived, the players arrived, and then the consumer digital camera arrived faster. PCD never made the jump to a world where images originated digitally rather than coming off scanned negatives. Kodak stopped extending the format; major applications quietly dropped full support. The files still exist in archives; reading them reliably today requires older software or specific plugins.
FPX — FlashPix
A coalition that included Kodak, Microsoft, Hewlett-Packard, and Live Picture developed FlashPix in 1996. It used a tiled, multi-resolution structure — conceptually similar to the way modern COG (Cloud-Optimized GeoTIFF) handles large raster data — so an application could load only the resolution tiles it needed rather than decoding a full image. The format embedded image processing instructions as a pipeline inside the file itself, so different applications could render the same FPX file differently depending on the end use.
FlashPix was supported in Internet Explorer through an ActiveX plugin and appeared in some digital cameras and consumer editing software around 1997–1999. Then nothing much happened. The multi-resolution tile approach was genuinely useful, but JPEG served the web well enough, professional tools preferred TIFF, and the format’s complexity worked against it. By 2001 it had largely disappeared from shipping products.
LWF — Lurawave
Lurawave used wavelet compression — similar in principle to JPEG 2000 — and was briefly positioned as a higher-quality alternative to JPEG for professional publishing in the late 1990s. Wavelet-based codecs can preserve fine detail better than DCT-based compression at equivalent file sizes because wavelets operate on the whole image rather than fixed 8×8 pixel blocks, so they don’t produce the block-boundary ringing that makes heavily compressed JPEGs look tiled. Lurawave had proprietary licensing, limited open-source support, and no browser adoption. It survived in some document management and aerial imaging niches before effectively disappearing.
JPEG 2000 — The Standard That Standardized Itself Into Irrelevance
JPEG 2000 warrants its own mention even though it achieved ISO standardization in 2000 and hasn’t technically “died.” It supports lossless and lossy compression, high bit depths, wide color gamuts, alpha transparency, and region-of-interest encoding. It is genuinely superior to baseline JPEG on several technical dimensions. Cinema DCP packages use a profile of it. Digital preservation workflows sometimes use it.
It never replaced JPEG on the web. Patents on the wavelet algorithms contributed to early hesitancy, browsers were slow to support it, and the computational cost of decoding was meaningfully higher than JPEG when it was introduced. Safari added support for .jp2 files at one point; that support has had a complicated history, and relying on JPEG 2000 for any general-audience web use has never been a safe assumption. Check current browser support tables before relying on it.
MNG — Multiple-image Network Graphics
MNG was developed alongside PNG in the late 1990s as a format for animated images — essentially an animated PNG, before APNG (Animated PNG) existed. It supported complex animation features including delta frames, object layering, and loops. It was technically more capable than animated GIF. Mozilla added MNG support in early Firefox versions, then removed it in 2004, citing the format’s complexity and limited adoption elsewhere. Without browser support, the format had no path forward. APNG eventually solved the animated-PNG problem with a much simpler specification.
JNG — JPEG Network Graphics
A companion to MNG, JNG was meant to bring JPEG compression to the PNG-style container, including alpha channel support — something baseline JPEG doesn’t have. A JPEG file stores only RGB values; it has no channel for transparency. JNG addressed this by embedding a JPEG-compressed RGB image and a separate grayscale PNG for the alpha mask inside the same container. The spec existed, implementations were partial, browser support never arrived, and it was superseded eventually by formats that handled transparency natively. It appears occasionally in game assets from the early 2000s.
XBM — X BitMap
XBM is old enough to feel archaeological. It was used in the X Window System, the graphical environment for Unix workstations, from the 1980s onward, and was actually a plain-text format — the image data is encoded as a C-style array of hexadecimal bytes that you can open in a text editor. Early web browsers including Mosaic and early Netscape supported it for small monochrome UI graphics. It is binary-free, purely human-readable, and completely unsuitable for anything beyond 1-bit monochrome icons. Modern browsers technically still parse XBM files in some implementations, though whether that persists is worth checking per browser version.
WBMP — Wireless Bitmap
Developed for early WAP (Wireless Application Protocol) mobile browsers around 1999, WBMP stored 1-bit monochrome images — pure black-and-white, no grayscale — with minimal header overhead, because the WAP stack was designed for phones with extremely limited bandwidth and memory. The format did exactly what was required of it in that context. Mobile browsers and screens changed; WBMP did not. It was listed in early HTML5 drafts and then dropped. Finding a device that renders it correctly today takes some effort.
PICT — Macintosh Picture
PICT was the native image and vector format on Classic Mac OS from the early 1980s through the late 1990s. It could contain bitmapped image data, vector drawing commands, and text in a single file — a mixed-mode container that predated PDF’s similar ambitions. Older Mac applications and clipart libraries are full of .pct files. Apple deprecated PICT when moving to macOS and the Quartz display system, which uses PDF as its underlying imaging model. PICT files can still be opened in some current software, but the format has had no development attention in decades. Its mixed-content structure makes it genuinely difficult to render correctly in modern contexts.
SWF Embedded Rasters
Technically SWF (Shockwave Flash) was never an image format, but it served as one in practice: enormous quantities of web graphics from roughly 1997 to 2015 existed only inside SWF containers, embedded as bitmaps compressed with the format’s own JPEG-variant or lossless ZLIB compression. When browsers dropped Flash plugin support — a process that completed across major browsers by the early 2020s — all of that imagery became inaccessible in a browser context without extraction tools. The pixels are recoverable from the files; the format they live in is no longer a viable delivery mechanism.
What Actually Kills a Format
Looking at these together, a few patterns emerge. Technical quality — better compression, richer feature sets, wider color support — does not reliably predict survival. PCD and FlashPix were technically capable. JPEG 2000 remains technically superior to baseline JPEG in several measurable ways. What mattered more was browser adoption, patent-free licensing, ecosystem timing, and whether the format solved a problem that content creators actually had at the moment they were choosing tools.
The sustainability of any format depends heavily on whether organizations with archival interests continue maintaining decoders. That concern — what happens to image data when its native software disappears — comes up directly in our coverage of file format sustainability and digital preservation, which is worth reading alongside this history.
For anyone managing legacy archives, the practical takeaway is specific: if you find PCD, FPX, or PICT files in an old collection, migrate them before the software capable of reading them becomes unavailable. That window narrows gradually, then quickly.