Why Patents Prevented JPEG 2000 Adoption on the Web

When JPEG 2000 was introduced at the turn of the millennium, it promised vastly superior compression, lossless scaling, and artifact-free imagery, yet it completely failed to become the universal image standard for the open web. The primary barrier was not technical capability, but rather the severe threat of patent litigation and legal ambiguity surrounding its underlying technologies. Fearing catastrophic intellectual property lawsuits similar to the Unisys GIF controversy, browser vendors and open-source developers refused to natively implement JPEG 2000, leaving the legacy 1992 JPEG format entrenched across the internet.

During the late 1990s, the web development ecosystem was severely disrupted by the enforcement of Unisys’s patent on the LZW compression algorithm used in GIF files. This dispute left software developers and web browser companies hyper-vigilant about intellectual property encumbrances. When the Joint Photographic Experts Group finalized JPEG 2000 (ISO/IEC 15444-1), they claimed that Part 1 of the standard could be implemented without paying royalties. However, the standard relied heavily on complex mathematical methods, notably the Discrete Wavelet Transform (DWT) and MQ arithmetic coding, both of which were surrounded by a minefield of proprietary patents held by various tech corporations.

The core issue was the danger of "submarine patents." Even though companies participating in the standardization process agreed to license their essential patents royalty-free, third parties who had no role in the committee retained patents that arguably covered the standard's implementation. In the proprietary software world, commercial entities could negotiate cross-licensing deals, but the open web relied heavily on open-source browsers like Mozilla Firefox. Open-source maintainers could not accept standards that might trigger royalties or infringement claims for end users and downstream distributors.

The legal ambiguity around arithmetic coding was particularly problematic. Arithmetic coding provided significantly better compression than traditional Huffman coding, but it had long been encumbered by patents owned by companies like IBM, AT&T, and Mitsubishi. While the baseline JPEG format had also included an arithmetic coding option, it was avoided for web use for the exact same reason: patent holders actively enforced their rights, forcing the web to stick with unpatented Huffman encoding. Browser developers were unwilling to risk baking arithmetic coding decoders into their software without ironclad guarantees of legal indemnity.

This licensing anxiety directly influenced major browser vendors. Microsoft, Mozilla, and eventually Google systematically refused to ship native JPEG 2000 decoders out of the box. Safari eventually added support within Apple's closed ecosystem, where the company already had enterprise patent protection, but without cross-browser consensus, web developers could not deploy the format publicly. Serving .jp2 files risked rendering blank spaces for the vast majority of web traffic.

The legal risks also exacerbated JPEG 2000's secondary technical drawbacks. Decoding wavelet-based images required significantly more processing power and memory than the original Discrete Cosine Transform (DCT) used by JPEG. For browser vendors, accepting elevated CPU consumption in exchange for a format that carried severe legal liabilities was an untenable trade-off.

Ultimately, JPEG 2000 found a secure home in specialized, closed industries where organizations could afford specialized licensing and high compute overhead, such as digital cinema (DCI), medical imaging (DICOM), and institutional archiving. For the open web, however, the fear of intellectual property liability ensured that JPEG 2000 was dead on arrival, leaving the door open decades later for explicitly royalty-free, modern alternatives like WebP and AVIF.