In our mission to make every stream look fantastic, we’re continuously pushing the boundaries on improving video quality. We currently offer the ability to optimize encoding based on the content of each video with Per-Title Encoding. Now, we’re excited to introduce Audience Adaptive Encoding, an encoding method that uses viewer data to deliver the best quality video tailored to your audience.
Audience Adaptive Encoding uses information on your audience bandwidth and device resolutions to provide a dynamic bitrate ladder that serves the highest quality video for your viewers. Based on your specific audience profile, we create renditions to match the distribution of bandwidths and resolutions so that more of your viewers get the best possible experience.
This new encoding feature is a part of our encoding process that uses machine learning using both content and context to give you the absolute best range of renditions for each and every video. It works concurrently with Per-Title encoding to bring the highest quality video to as many of your viewers as possible.
For a high-level look at how this works, let’s look at viewer distribution for bandwidth. In this example, based on an analysis of this particular video, our Per-Title system determines that the optimal top bitrate is 5Mbps (the blue line). From there, we determine additional renditions that will maximize video quality for the largest number of users (the red lines).
Here, you’d get renditions for 1.5Mbps and 3.2Mbps. By providing a rendition that fits your viewer’s bandwidth, video quality increases while rebuffering and startup time decreases.
We’ll share more details on how Audience Adaptive Encoding works. Keep an eye out for a follow-up post.
Good news! If you use Mux Video, you’ll soon be getting Audience Adaptive Encoding. (It’s currently available behind a feature flag while we roll it out to all users.) However, in order to fully maximize performance, you’ll need to also integrate with Mux Data. If you use Mux Video and: