Mark Wachsler is a senior software engineer on YouTube's Video Infrastructure team, focusing on video transcoding and video quality. Prior to Google he had 20 years of experience designing audio and video algorithms, software, and hardware at music technology company MOTU. Mark earned a master's degree in Computer Science and a bachelor's degree in Cognitive Science at Brown University.
This talk covers some of the research which has been done by YouTube's Media Algorithms team. The Rate Quality Optimization algorithm was developed to reduce streaming bitrate while preserving visual quality. The talk will explain how it has been deployed to improve QoE for YouTube's most popular content. On the other hand, we understand that User Generated Content is very different from traditional video and should be treated as such. To assist in the advancement of video compression and quality assessment research of UGC videos, we released the UGC dataset, a sampling of UGC uploaded to YouTube. The talk will then describe the methodology used to create the dataset, and deal with some of the challenges involved in processing UGC.