How Google Built JPEG XL: A New Image Format From Open Source Research
Google has published a detailed account of how open source projects led to the creation of JPEG XL, a new image format now adopted as an international standard. The company's path to JPEG XL began with engineers trying to squeeze more efficiency out of the old JPEG format, then progressed through two research projects called Guetzli and Brunsli, and eventually led to an experimental codec named PIK. Two of PIK's core techniques ended up in the final JPEG XL standard (formally known as ISO/IEC 18181).
From Fixing Old JPEG to Building Something New
Google's strategy started practical: extract as much compression as possible from the decades-old JPEG format while keeping it compatible with existing software. Guetzli optimized JPEG by analyzing which image details the human eye actually notices, improving the compression ratio without making photos look worse. Brunsli took a different angle, using advanced mathematics (a technique called entropy coding) to shrink file sizes while keeping the files readable by ordinary JPEG software.
The research teams then released PIK as an open source project. PIK combined the best ideas from both earlier projects — Brunsli's compression techniques and Guetzli's focus on visual quality. Later, PIK's algorithms were merged with similar work from another company (Cloudinary) that had developed a competing approach called FUIF. The result was a single standard that pooled the strongest techniques from multiple independent research efforts.
This pattern of competing projects merging into one agreed-upon standard is not new in image technology. When the web transitioned from GIF (an older image format) to PNG in the 1990s, many companies experimented with different compression methods before settling on a single open standard that offered the right balance of efficiency, visual quality, and ease of implementation.
What JPEG XL Actually Does
JPEG XL shrinks image files to about 40% of their original size while keeping the visual quality identical to the old JPEG format. It also adds modern features that older JPEG cannot handle: high dynamic range (better contrast and color), animation (like GIF), transparency (the see-through parts of an image), and both lossy and lossless encoding modes (meaning you can choose between smaller files or perfect quality).
A particularly useful feature: servers can store one file and serve it to different types of devices. The new JPEG XL standard includes techniques to convert files efficiently, so users with older browsers get traditional JPEG while newer browsers get the smaller JPEG XL version.
The standard was designed to work with regular software running on ordinary computers, not specialized hardware. This matters more than it might sound — older image formats sometimes required expensive dedicated chips to encode or decode, which slowed their adoption in real-world systems.
Google also released a tool called Jpegli, a software library that borrows compression techniques from JPEG XL and applies them to old-style JPEG files, achieving about 35% better file compression for high-quality images. This gives companies a path forward: they can improve their existing JPEG files without immediately switching to the new format.
The Standardization Work and What's Available
The JPEG Committee officially adopted JPEG XL as an international standard. Google released the reference implementation (libjxl) as open source software that anyone can use. The format was designed for web environments where bandwidth matters and images need to display correctly on phones, tablets, and desktops.
During development, Google's own teams tested JPEG XL on their image storage systems to see how it performed in the real world. They also released additional tools, including machine learning software designed to help with JPEG XL's more advanced compression features.
The broader context here is that JPEG XL is the first major image format redesign since WebP came out over a decade ago. While WebP was primarily about faster downloads, JPEG XL also focuses on features that matter for modern photography: very high contrast images, animation, and images created by computational photography (advanced camera software that processes multiple exposures).
The Adoption Problem
Technical quality does not automatically translate to adoption. Google faced criticism when it decided not to support JPEG XL in the Chrome browser, with the Free Software Foundation and other groups objecting to the decision. This highlights a real tension: standards need broad platform support to succeed, and browser makers do not always prioritize the same formats.
History shows this pattern. WebP, which was superior to older formats, took many years to gain browser support, and AVIF (another newer format) still has incomplete adoption even today. When a company decides whether to add support for a new image format, they must weigh the technical benefits against the effort required and how many users actually have software that supports it.
JPEG XL addresses one traditional obstacle to format adoption. Its ability to convert existing JPEG files without re-encoding the entire library is a practical advantage. Companies do not have to rebuild their entire archive of images — they can transition gradually.
The next year or so will likely settle the question of whether JPEG XL becomes widely used or remains a technically sound but underutilized option, like other formats that failed to overcome deployment obstacles. Key signs will be whether major companies adopt it and whether browser vendors reconsider their support for it.
