Centered Sample-Range Remapping at JPEG XL Effort 6

This preprint evaluates a reversible sample mapping that embeds 8-bit RGB values in the central half of a 9-bit sRGB code space before lossy JPEG XL encoding. The stored file includes a five-byte header that identifies the inverse mapping. The evaluation covers 41 previously unused CLIC Professional and DIV2K validation images at their original resolutions. Under common RGB PSNR and MS-SSIM floors and a single-thread encoding cap of 0.82 seconds per megapixel, the centered mapping reduced aggregate file size by 6.71% relative to direct JPEG XL at effort 6 (95% bootstrap interval: 3.85% to 9.86%). The centered files were 3.94% larger in aggregate than the tested AVIF files. Same-depth and transfer-function controls, together with libjxl coding counters, examine how the input representation affects the result.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24207466
Primary Topic
Advanced Data Compression Techniques
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Centered Sample-Range Remapping at JPEG XL Effort 6

Ali Ghanbari
Zenodo (CERN European Organization for Nuclear Research)
Advanced Data Compression Techniques
preprint

Centered Sample-Range Remapping at JPEG XL Effort 6

Ali Ghanbari
preprint en

Abstract

This preprint evaluates a reversible sample mapping that embeds 8-bit RGB values in the central half of a 9-bit sRGB code space before lossy JPEG XL encoding. The stored file includes a five-byte header that identifies the inverse mapping. The evaluation covers 41 previously unused CLIC Professional and DIV2K validation images at their original resolutions. Under common RGB PSNR and MS-SSIM floors and a single-thread encoding cap of 0.82 seconds per megapixel, the centered mapping reduced aggregate file size by 6.71% relative to direct JPEG XL at effort 6 (95% bootstrap interval: 3.85% to 9.86%). The centered files were 3.94% larger in aggregate than the tested AVIF files. Same-depth and transfer-function controls, together with libjxl coding counters, examine how the input representation affects the result.

Zenodo (CERN European Organization for Nuclear Research)
Advanced Data Compression Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon