Overview: The Challenge
By 2014, all forms of video content will account for 91% of global consumer traffic, generating 50 exabytes of data per month (Cisco, 2010)
Global IP traffic is expected to grow by 34% per year through 2014, reaching 767 exabytes per year or 64 exabytes per month in 2014 (Cisco 2010)
In July 2010, 178 million Americans watched online video content, resulting in 5.2 billion video sessions and an average of 14.7 hours per viewer. The average length of each video = 4.8 minutes (comScore)
YouTube serves over 2 billion videos daily worldwide (YouTube, 2010)
Investment in new internet infrastructure in the US alone will exceed $100 billion in 2012… due largely to the exploding demand for video-based internet services (Nemertes report, 2007)
From Macroblock to Computer Vision, Image Understanding, and Pattern Recognition
Video demand is exploding. File sizes are increasing. People expect high definition quality.
Video – anytime, anywhere.
The current paradigm of utilizing block-based video coding has successfully been extended over several generations of encoding technologies and, according to many experts, is considered to be nearing the end of its life cycle, having delivered most of its achievable compression benefits. There are still many beneficial enhancements that can be added to this existing technology through enhanced encoders and new expansions to the standard. While these steps will provide a consistent and modest gain in performance over time, there is general agreement that block-based video coding will require the addition of higher level modeling to achieve the next big leap in compression.
As video demand continues to grow and as video file sizes continue to expand (from Internet Video to Standard TV to High Definition resolutions and beyond), new means of achieving additional compression – without losing visual quality – moves from an attractive proposition to a requirement.
Euclid Discoveries has been pursuing the goal of identifying and implementing a new compression technology since 2003 and is dedicated to creating the key enabling parts of the next generation in video compression. The company’s research and development has focused on combining video compression algorithms with techniques originally developed for Computer Vision, Image Understanding and Pattern Recognition. The result of these efforts is eFLEX, Euclid’s prototype implementation of the next generation encoder technology based predominantly on existing H.264 algorithmic elements.