Publications / Ready, steady, content

Microsoft

Microsoft News is a comprehensive digital platform providing up-to-the-minute news updates from around the world. Leveraging reliable sources and AI technology, it delivers a wide array of stories across diverse categories, making it a one-stop hub for global insights and information.


Thursday, 08 November 2018

Speaking at the Media Hack Weekend in 2015, I delivered the following text to the epigraph of my presentation: "Cowering like a spring, the runner quickly throws the body forward after a shot. To do this, he usually uses special starting blocks, which are installed in front of the line beyond which running distances begin. They give the runner a solid footing for instant, powerful repulsion." Today cloud services are these starting blocks.

Global media industry experience

For many years I am observing the evolution of the IT industry. Every year integrated solutions are becoming more comfortable and cheaper. Today you need to choose the right components from which you can build the system. Most of the projects with which I have worked for a long time are somehow explicitly related to the field of media. In each of the projects, there were many tasks that ten years ago would put me in a rather tricky situation — I would either have to rent expensive equipment or build my own data center. Fortunately, cloud solutions already available today. For each of the tasks, now it is possible to choose several alternative technologies from different vendors. In most of my projects, I used services based on Microsoft Azure, so most of this platform's examples are examples.

Look at those projects and companies that have already managed to implement cloud technologies in their products and projects:

Broadcast of the Olympics in 2012 in London.The capabilities of Azure Media Services were used to deliver real-time and on-demand video streams to several broadcasters that broadcast the Olympic Games. Among them: France Télévisions, RTVE (Spain), CTV (Canada), and Terra (Central and South America). In collaboration with Deltatre, Southworks, Gskinner, and Akamai, over 2,300 hours of live broadcasts and HD content from the Olympic Games were delivered to more than 20 countries.

Agency Reuters. One of the world's largest international news and financial information agencies uses the capabilities of Microsoft Azure Machine Learning to implement a video selection mechanism. When viewing a page dedicated to an event, the most relevant video is automatically searched.

Nasdaq Tower . For nearly 20 years, the impressive Nasdaq MarketSite video tower has illuminated Times Square in New York. It has become a landmark attraction and a popular platform for live broadcasts, vibrant advertisements, and market quotes' rapid publication.

Moving the infrastructure for working with video, graphics, and visualization to Azure will allow Nasdaq to scale workflows quickly. Nasdaq is also exploring other advanced Azure features, such as machine learning and artificial intelligence, to differentiate work with MarketSite from anywhere in the world.

Battle for attention

It's no secret that when working with media projects, there is a need to process and store files of a rather large size. So, selecting raw RAW photos or an unprocessed video file can take up to several tens of gigabytes. And after processing, the material ready for broadcast and publication can take up a little less space. Do not forget about the widespread distribution of devices with high-resolution support, which require the very same high-definition video files by no means small. The Azure Storage service, in this case, will perfectly cope with the task of storing all the information. At the same time, the cost of a cloud solution will be quite affordable.

The next task, which usually follows the task of storing data, is the task of delivering media files. If you have an Internet TV channel that is watching by viewers around the world, you have to think about which content delivery networks to use and how to integrate with them. In this case, there is nothing more convenient than the built-in tools for working with the Content Delivery Network in Microsoft Azure. You not only get full control over your content, but you can also choose which provider use: Akamai, Verizon, or the servers of Microsoft itself.

In addition to tasks related to the content, the trend towards implementing feedback and interactivity is gaining popularity in modern media. be almost wholly transferred to the cloud platform, then at the stage of creating the web application, you should connect a professional development team that can deal with the software part of the project. At the same time, however, do not forget about the crucial specifics of media projects — this is the presence of peak loads. For example, if once a week you have a new release, and even more so, if it is a live broadcast and the audience is also a participant in the event, then at this moment, the load on your entire infrastructure can increase hundreds or even thousands of times. To effectively handle peak loads and not overpay for server resources that will remain idle the rest of the time, you can use Azure Web Apps, a platform for creating and deploying web applications. Web Apps services allow you to adjust the scaling of server resources, which can help ensure the system is operational at peak loads while consuming only the necessary amount of capacity the rest of the time.

AI in the media

Separately, it is worth noting the development of services based on machine learning that allows you to implement such scenarios of using work that previously we can only dream of. Here is just an incomplete list of what a video indexer from the Azure services suite allows you to do:

  • Automatic video language recognition;
  • Determination of the number of speaking persons;
  • Recognition of the text shown on the video;
  • Key personnel extraction;
  • Analysis of tonality — the definition of positive, negative, and neutral tonality in speech and visual text;
  • Moderation of visual content;
  • Extracting keywords from speech;
  • Definition of trademarks;
  • Celebrity identification;
  • Description of obscene text in speech decoding;
  • Recognition of plots;
  • Recognition of emotions;
  • Translation of speech transcript into 54 different languages.

A few years ago, the implementation of almost each of these opportunities would require either the development of their complex algorithms or the hiring of employees who would manually analyze video content. Today, it's all available in a video of simple and convenient services and APIs that even one developer can handle.

Modern consumers of content force the media to change, be flexible, use modern solutions, and in practice, follow the ideas of digital transformation. Oddly enough, it is appropriate to recall the epigraph about the starting blocks here again. This text was published in the children's magazine "Bonfire" in 1968, half a century ago. And today, this content continues to live on the Internet, and many copies are in various cloud storage, which allows it to find its reader and remain relevant.


Read publication