Digital archivists, photographers, and data hoarders routinely battle storage constraints. JPEG XL—a modern image format developed by the Joint Photographic Experts Group—offers a compelling solution: lossless JPEG transcoding that shrinks file sizes by an average of 22% without altering original pixel data. Unlike traditional lossy conversions (e.g., JPEG to AVIF), this technique repackages existing JPEG information into JPEG XL's efficient container, enabling bit-for-bit reconstruction of the source file.

The Technical Workflow

Conversion relies on the reference implementation's cjxl tool. For lossless transcoding:

cjxl -j 1 -e 10 -d 0 image.jpg image.jxl
  • -j 1: Enables JPEG reconstruction mode
  • -e 10: Maximizes compression effort (trades CPU time for smaller output)
  • -d 0: Ensures lossless handling if source isn’t JPEG

Verification uses djxl and cmp:

djxl image.jxl decoded-image.jpg
cmp image.jpg decoded-image.jpg || echo "Reconstruction failed"

Batch processing leverages find and GNU Parallel:

find /gallery \( -name \*.jpg -o -name \*.jpeg \) -print0 |
  parallel --null nice python3 jpg-to-jxl.py

Limitations and Ecosystem Gaps

Despite efficiency gains, JPEG XL faces adoption barriers:
- Limited native support: Safari offers basic decoding; Chrome removed experimental support. Web deployment requires on-the-fly conversion back to JPEG for compatibility.
- Tooling inconsistencies: Applications like Geeqie may mishandle embedded color profiles. Malformed JPEGs might require --allow_jpeg_reconstruction 0, sacrificing bit-for-bit fidelity.
- Opaque compression metadata: JPEG XL doesn’t flag whether content originated from lossy/lossless sources—a concern for preservationists.

Lossless Compression Benchmarks

For non-JPEG originals, JPEG XL’s lossless mode competes with WebP and PNG:

Format Total Size Savings vs PNG
PNG 9.60 GB —
WebP 7.63 GB ~20%
JPEG XL 6.39 GB ~33%

Results from 1,346-image test corpus.

Strategic Implementation

Author Michał Nazarewicz recommends a hybrid approach: Transcode JPEGs to JPEG XL for storage savings, but convert lossless originals (like PNG) to WebP—which preserves compression-type metadata. While JPEG XL’s space efficiency is transformative for storage-heavy use cases, its ecosystem maturity lags behind its technical promise. Widespread adoption hinges on broader toolchain integration and browser support.

Source: mina86.com