Netflix is making significant strides in advanced video encoding, thanks to the efforts of Anne Aaron and her team. Over the past 13 years, Aaron has been optimizing the way Netflix encodes its movies and TV shows, resulting in better-looking streams over slower connections and a 50% bandwidth savings for 4K streams alone.
One of the key technologies behind this optimization is Video Multi-method Assessment Fusion (VMAF), a video quality metric that combines human vision modeling with machine learning. VMAF was born out of a research collaboration between Netflix and Prof. C.-C. Jay Kuo from the University of Southern California. The goal was to develop an automated way to answer the question, “How will a Netflix member rate the quality of this encode?”
VMAF has proven to be a valuable tool for quality assessment, with its adoption extending beyond Netflix to the broader video community. It has been integrated into third-party video analysis tools and used in industry trade shows and meet-ups to compare the quality and efficiency of various encoding techniques.
This advanced encoding strategy allows Netflix to deliver higher quality video under low-bandwidth conditions. For instance, titles with “simple” content can now be streamed at a higher resolution for the same bitrate.
Read more: www.theverge.com