At the 2017 Mobile World Congress show in Barcelona, Netflix discussed the advancements in technology regarding their platform over the past year. Among the things mentioned were an increase in binge-watching and mobile phones being used more and more to consume Netflix-content. Instead of designing mobile-centric videos, Reed Hastings and the team at Netflix decided to focus on creating better codecs that can benefit all users of the service. This is the idea that birthed the Dynamic Optimizer AI.
Unlike normal video compression methods, the Dynamic Optimizer can optimize particular scenes in the streaming content in real-time. This works best for viewing Netflix in low bandwidth environments (such as mobile streaming or network congestion at home), giving more bandwidth to action-packed scenes and less bandwidth to talking heads or monologs for example. The end result of the Dynamic Optimizer is that all users should get less buffering and interruptions on mobile and non-mobile devices alike.
Netflix worked with the University of Southern California, and the University of Nantes to create the Dynamic Optimizer AI. The algorithm powering this AI was trained by the combined ratings of hundreds of viewers judging the quality of hundreds of thousands still shots. The purpose of the Dynamic Optimizer is to provide the same visual fidelity across all screen types and networks, especially with those networks with a low bandwidth limit. The Vice President of Innovation at Netflix, Todd Yellin stated that a lot of users in India, Japan and South Korea watch Netflix on their mobile devices so it is important that they experience the content the same as others. The Dynamic Optimizer should be rolling out to in a couple of months, much to the delight of binge-watchers everywhere.