The project aims to visualize the bitrate and encoding speed of libjxl with data following the benchmark_xl format.

Read, watch or cite our paper

1. Install the dependencies with pip install -r requirements.txt

Generate the primary data using bash src/generate_primary_data.sh <path to the images directory> or using the benchmark_xl tool with the --print_details_csv option.

⚠️ Please note that benchmark_xl yields different results than cjxl.

The following script will produce a graphic of the bitrate according to the quality factor and effort. The API of the script can be found using python src/visualize_bpp.py -h.

The following script will produce a table of the encoding speed according to the quality factor and effort. The API of the script can be found using python src/table_enc_speed.py -h.