Researchers Release New Algorithm to Enhance Streaming Video Technology

The algorithm was created with the aid of volunteer spectators who viewed a stream of footage, served by computer scientists who used machine learning to track this data flow in real-time, searching for ways to minimize glitches and stalls.

FREMONT, CA: Researchers have released the latest algorithm that shows a significant advancement in streaming video technology.  The algorithm was created with the aid of volunteer users who watched video feeds, assisted by computer scientists who used machine learning to track this data flow in real-time, searching for ways to minimize glitches and stalls.

In a study, the researchers explained how they created an algorithm that only produces as much data as the viewer’s Internet connection can obtain without degrading quality. In streaming, evading stalls depends heavily on these algorithms. The study showed that there might be a gulf in making one of these algorithms work in simulation and making it work on the real internet.

According to the researchers, many of the prevalent video streaming solutions are based on something called the Buffer-Based Algorithm, known as the BBA. BBA essentially asks the viewer’s device how much video it has in its buffer. For instance, if it is stored for less than 5 seconds, the algorithm sends poorer quality footage to protect against interruptions. If the buffer is stored for more than 15 seconds, the algorithm sends the best possible video quality. If the number collapses in between, the algorithm modifies the quality accordingly.

While BBA and related algorithms are common in the industry, there have been constant attempts by researchers over the years to create more complex algorithms using machine learning—a method of Artificial Intelligence (AI) in which computers teach themselves how to optimize certain processes. Generally, these machine learning algorithms need simulated data to learn from rather than the actual thing delivered over the real internet. Therein lies a concern, according to the researchers.

Check This Outlook : Top Media and Entertainment Solution Companies