The JPEG XS (ISO/IEC 21122) standard defines a compression algorithm with very low latency and very low complexity. By offering various degrees of parallelism, JPEG XS can be efficiently implemented on various platforms such as FPGAs, ASICs, CPUs and GPUs and excels with high multi-generation robustness. It is particularly optimized for visual lossless compression as defined in ISO/IEC 29170-2 for both natural and synthetic images. The typical compression ratios are between 1:2 and 1:6 for both 4:4:4 and 4:2:2 images and image sequences with up to 16 bit component precision. Typical parameterizations address a maximum algorithmic latency between 1 - 32 video lines for a combined encoder-decoder suite.
Those properties allow JPEG XS to be applied in all applications that before operated with uncompressed image data. This includes among others professional video links (3G/6G/12G-SDI), IP transport (SMPTE 2022-5/6 & SMPTE 2110), real-time video storage, memory buffers, omni-directional video capture system, head-mounted displays for Virtual or Augmented Reality and sensor compression for the Automotive Industry. By these means, systems can offer an increased resolution and frame rate, while offering a visually lossless quality.