Disney+ Hotstar Introduces New AI Video Optimisation Tool to Offer Data-Efficient Viewing Experience

Disney+ Hotstar, the over-the-top (OTT) platform, introduced a new artificial intelligence (AI) tool on Monday that will improve the user experience. The streaming giant claimed that the AI tool will optimise video streaming so that less data is consumed while keeping the quality of the video unchanged. The new feature will also make it more data-efficient to download videos from the platform for offline viewing. Notably, the company announced last month that it would stream a select number of English Premier League matches in 4K resolution.

Disney+ Hotstar to Use an AI Video Optimisation Tool

In a press release, the video-streaming giant announced the introduction of AI-powered video optimisation technology. It is essentially an advanced encoding feature that can optimise data consumption based on user habits and scene complexity. Disney+ Hotstar claimed that the technology will let users spend less data while keeping the same quality of visuals.

It also added that users can expect a reduction of 25 percent when downloading videos for offline viewing without losing quality. This is aimed at making the content on the platform more accessible to users with limited data.

Detailing the AI feature, Disney+ Hotstar said that the technology uses AI to understand user consumption patterns as well as the context and complexity around each scene. While it is difficult to understand how user patterns are linked with data optimisation, the context and complexity of the scene are likely being processed with computer vision to brighten darker scenes while not increasing data usage. The company highlighted that the technology is currently pending patent.

Mukund Acharya, the Head of Technology at Disney+ Hotstar further explained that the new AI-powered encoding technique can improve the watch time per GB by 12.5 percent. This means users can watch more content on the same data budget.

The company also mentioned that the feature was developed using various AI-based models, however, it did not specify whether these were native models or third-party models.