51 YOUR TAKE AI
every new video file. With AI, algorithms can seamlessly run and identify encoding parameters based on variables, like screen size, that can have a major impact on video quality – and ultimately the end user experience, resulting in faster processing times and significantly higher quality with no increase in bandwidth. Added to this, machine learning enables a continuous learning process, which brings with it enormous benefits for encoding. Every encode is absorbed and remembered by the AI functionality, enabling it to make the predictions for future encodings more accurate. This not only increases the quality of experience and quality of service for end users, but also reduces the number of bits required to achieve the same quality stream. AI FOR AUDIENCES Beyond encoding, AI also has the power to drive an enhanced user experience
by intuitively applying key image tools such as facial and speech recognition. If a user were to fall asleep or leave the room, for instance, the AI functionality could enable the video to be paused and then resume once the viewer looked back at the device. AI could also enable the volume of a video to be reduced if a user were to answer a phone call, and then return to the original volume once the call has ended. When it comes to content organisation, AI can create rich metadata tags for each piece of content, enabling the industry to monetise those assets more effectively. Furthermore, AI can reduce engineering costs by automating any number of tasks, enabling technical staff to allocate their time to other important areas. AI algorithms are already in use by streaming services such as Netflix and YouTube to address latency issues, encoding costs and customer churn.
Netflix estimates that its use of AI to automate workflows and reduce customer churn saves the company around $1 billion annually. The streaming provider uses AI to recommend better content depending on viewing preferences, as well as to optimise video compression and delivery. IMPROVING USER EXPERIENCE Although the industry is still in the early stages of the AI adoption curve, some benefits are already evident. There is enormous potential for AI-based models to help predict implications to video workflows when changing certain settings, like bitrate ladders, without the need for testing. Such approaches will enable TV and media companies to maximise the value of their services by reducing the complexity of their workflows and delivering high-quality content to any device. Simply put, AI makes video encoding more dynamic and improves the user experience. Due to the transformative power that we have seen demonstrated in the wider video distribution ecosystem, namely in enhancing efficiency and cutting down operational costs, we expect new AI-enabled encoding, playback and analytics developments to start hitting the market in the coming months.
MACHINE LEARNING ENABLES A CONTINUOUS LEARNING PROCESS, WHICH BRINGS WITH IT ENORMOUS BENEFITS FOR ENCODING
feedzine feed.zine feedmagazine.tv
Powered by FlippingBook