Fraunhofer-GesellschaftVideo streaming content differs in terms of complexity and requires specific encoding settings to achieve a certain visual quality. Our Deep Encode solution leverages machine learning techniques to avoid test encodes and enables per-scene encoding for live linear content. We automatically predict video quality metrics like PSNR, SSIM and VMAF, detect complex metadata like scene changes and perform object classification and labeling for each video to provide the optimal encoding ladder for your applications and services.
AI powered per-scene live encoding
in cooperation with
