Migration of the broadcast sector to the cloud
Consumer viewing habits and media business models have changed in recent years alongside the progress seen by smartphone technology and connectivity coupled with the mushrooming of over-the-top (OTT) services. In order to keep up with the growing demand for content and adapt to rapidly changing viewer behavior, mass media are turning to the cloud to reinvent the media chain, from acquisition to delivery of first-class IP, adaptive live bit rate and video on demand across multiple devices. In line with this trend, broadcasters are working to refine channels and offload a better part of their traditional video infrastructure onto the cloud. In doing so, the flexibility, scalability, and security provided by the cloud are proving to be a great asset and opening up new technological and operational freedom of action, as well as resulting in cost savings. Although a large majority of broadcasters have already started adopting cloud-based technology for these reasons, conducting a complete broadcast workflow in the cloud is becoming a reality thanks to technological advances that facilitate the delivery of Traditional broadcast content and multi-screen linear video within a unified virtual architecture.
In recent years, we have begun to see that more broadcasters prefer cloud-based resources for processing, packaging, and delivering live and on-demand video over traditional on-premises hardware, especially since cloud-based resources have become more affordable. This is in part because moving live linear workflows on to the cloud has significant benefits for remote production and collaboration, can greatly accelerate content production and delivery, and provides a platform for launching innovation in a faster way. By using the cloud, all of these benefits can be attained with lower initial and operational costs.
An example of this is the case of the FOX network, which has opted for AWS cloud services on all its channels with the aim of achieving a more efficient distribution of its sports, news, and entertainment television content to multi-channel video programming distributors to more than 200 affiliate stations and over-the-top (OTT) providers. FOX configuration readiness allows your team to react faster to market changes, while making content more accessible to a wider audience.
As cloud adoption by broadcasters such as FOX and others increases, technological innovation accelerates, laying the groundwork for the broadcasting industry to move the entire video channel to the cloud and gain significant efficiency along the process.
As in other industries, in the audiovisual creation and distribution sector we are also seeing strong growth in the use of Machine Learning and Artificial Intelligence, as it allows these companies to simplify many tasks.
For example, broadcast and live stream monitoring (OTT) service providers perform a large number of quality control checks in order to avoid errors. These range from low-level signal errors to high-level issues such as content errors. Traditional live media analysis software focuses on signal-level quality checks. High-level quality checks, such as verification of program content, subtitles, or audio language, are performed by human operators who constantly monitor the broadcast stream for problems. As the number of video streams broadcast increases, it becomes difficult and costly to expand the manual monitoring effort to support more channels and programs.
The latest advances in artificial intelligence can automate many of the high-level monitoring tasks that were previously fully manual. With the help of AI-based detections, human operators can focus on higher-level tasks, respond to problems faster, and monitor more channels with higher quality. Amazon Rekognition uses AWS AI services to analyze the content of a live video stream and is capable of performing a set of examples of near real-time monitoring checks.
However, the mass media have not only used these technologies for internal organization purposes, but also in their broadcasts. An example of this is the one Sky News carried out during the wedding of Prince Harry and Meghan Markle.
Unlike sporting events, whose broadcasting rights are subject to licenses that have been carefully granted for each country, a public show like a royal wedding allows multiple broadcasters to compete for a higher audience share, causing each broadcaster to offer more. value in terms of both quality and innovation.
For Harry and Meghan’s wedding, Sky News harnessed the power of AWS artificial intelligence to create “Sky News Royal Wedding: Who’s Who”, a live streaming application, based on machine learning, that allowed viewers to identify the wedding guests and participants in real time.
The live streaming experience allowed users to watch the arrival of guests up close, with the machine learning infrastructure to detect attendees in real time as they appeared on the screen. Half a million people used this feature in more than 200 countries through mobile devices and web browsers.
Sky News designed and created the machine learning functionality in just three months. The team built the service using cloud infrastructure for speed and growth, as well as to ensure access to scalable machine learning capabilities that are critical for the core functionality of the real-time guest identification application.
The project combined AWS cloud video infrastructure with GrayMeta’s data analytics platform and UI Centric’s interactive user interface design. In this case, the Amazon Rekognition image and video analysis service was also used to identify celebrities in real time and tag metadata with related information, integrating the user experience in collaboration with UI Centric.
In addition, in AWS we have some clients that use Cloud technology applied to sports broadcasts, such as the Bundesliga, Formula 1 or the NFL, among others.
For some time now, AWS has been working together with the Bundesliga to offer viewers -through Cloud technology- a deeper and more analytical vision of what is happening on the pitch. In fact, just a few weeks ago, AWS and the Bundesliga announced three new match-related pieces of data (Match Facts).
Match Facts are generated by collecting and analyzing data from live match video streams as they are transmitted to AWS. Fans will see this information in the form of graphics during broadcasts and on the official Bundesliga app during the 2020/2021 season and beyond. These advanced statistics help the public better understand things like the strategy involved in making decisions on the field and the probability of a goal per shot. These three new game data pieces better show the action on the field and provide spectators, coaches, players, and commentators with visual support to analyze a team’s decision-making.
The growth of advanced statistics in sports has led to a deeper understanding of strategies in the game and a better comprehension of the athletes’ performance.
The new match data: Player Most Pressured, which highlights how often a player experiences significant pressure during a match; Attack Zones, showing fans where their favorite team is attacking and from which side of the field they are most likely to score; and Average Positions – Trends, which shows how changes in a team’s tactical formation can affect the outcome of a match, join those already known: Speed Alert, Average Positions and xGoals.
Thus, Bundesliga fans already have 6 Match Facts that enhance their match viewing experience.
Another of the sports broadcasts that AWS works with is Formula 1. As we all know, F1 is a battle between the best drivers in the world, but it is also a battle between some of the most innovative engineers. No other sport has been so dynamic in its evolution and adoption of new technologies. That is why AWS is proud to be the official cloud and machine learning provider for Formula 1.
Furthermore, Formula 1 has been working with AWS for some years to create the statistics system based on artificial intelligence and data analytics, “F1 Insights powered by AWS”. This system analyzes a history of 70 years of data with information collected in real time to create a series of statistics on the performance of drivers and cars on the track. This technological system of statistics helps, on the one hand, riders, and teams since it gives them a new vision and data on their work; and on the other, it allows providing both F1 fans and spectators with a new experience to better understand what is happening on the track during broadcasts. At present, data are being provided on: Starting Speed, Predicted Pit Stop Strategy, Pit Scenario, Battle Forecast, Pit Battle Strategy, and Tire Performance.
Last, it is worth highlighting the case of the NFL, who chose AWS to make the most of all its data through sophisticated machine learning analysis. The NFL uses the power of AWS to create new statistics on game broadcasts and improve player health and safety by predicting potential injuries, thus creating a new experience for fans, players, and teams, all this in real time.
The NFL has created a number of new AWS-based statistics, each based on different data points and broadcast during game broadcasts. The most important among such functions are: Expected running yards, Expected yards after catching the ball and Route classification.
By Carlos Sanchiz, Manager Solutions Architecture at Amazon Web Services