MIT Developed Pensieve help Reduce video Rebuffering

In these times, immersed in an almost absolute hyper-connectivity, there is nothing more exasperating when surfing the internet than to “hang up” a video when it is being played. It’s frustrating at the moment when, for example on the YouTube network or Netflix-type services, the video suddenly becomes pixilated or stops altogether while trying to load content through online platforms.

It is the criticized system of “buffering,” but may have the days counted. The reason is that of the current algorithms used in digital platforms and consist of dividing the content into portions of videos that are uploaded as they are reproduced.

If you have a slow connection, a service like YouTube can reduce the image quality so that the image transfer is uninterrupted. But sometimes, it is the reproduction system itself that fails at the moment when it tries to skip to a part of the video that had not previously been uploaded, thus stopping the content with the frustration of the user.

To avoid this, a group of researchers at the Massachusetts Institute of Technology (MIT) has developed a technology based on Artificial Intelligence and using machine learning techniques, called Pensieve, which analyzes in real time the Connection conditions to show the quality of streaming video with a lower buffering load. Specifically, up to 30% less.

For this, it uses that algorithm known as ABR (Adaptative Bitrate) to determine the ideal quality that can be transferred through the network. And by deploying it, it has been proven to offer a higher quality streaming experience with less rebuffering than existing systems. This technology, in turn, can have a great impact not only in conventional video reproductions, but also in the use of content destined to virtual reality.

The idea is to adjust in real time the fragments of the video according to the forecasts of change of the type of speed of the connection in a way that reduces the chances of stopping. “Studies show that users leave video sessions if the quality is too low, which causes huge losses in advertising revenue for content providers, ” MIT professor Mohammad Alizadeh said in a statement. ‘Platforms must constantly seek new ways of innovating.’

This system, according to its developers, has an added advantage, flexibility. It can also be customized according to the priorities of an audiovisual content provider. For example, if a user goes through a tunnel where connection losses are recorded, a platform like YouTube could reduce the bit rate to load the video content needed and avoid cuts.

No ratings yet.

Please rate this