Can Artificial Intelligence improve the process of video compression? Attend this BEIT session and find out.
With digital content sources and resolutions expanding regularly, we’re going to need Artificial Intelligence to help us wrangle the data.
In the more than two decades broadcasters have been compressing digital video, there have been many improvements to the technology. Now, Artificial Intelligence (AI) is pointing the way toward ever more effective methods of squeezing the growing resolutions for broadcast and OTT content delivery into increasingly more efficient data payloads.
Jean-Louis Diascorn, senior product manager at Harmonic, and will present his paper, “AI Technology is Changing the Future of Video Compression,” on Monday, April 8th at 10:40AM session in room N256, as part of the Broadcast Engineering and Information Technology (BEIT) conference.
Diascorn is one of the leaders in this field, having worked on (among any other projects) some of the first SD and HD MPEG-2 encoders ever developed.
“This is really a revolution,” Diascorn said, “and to put it in context I plan to start with a brief overview of the history of video compression. Its first use for distribution started in the mid-1990s including multipass look-ahead encoding in 1998, and the addition of a Motion Compensated Temporal Filter (MCTF) in 2000. Then in 2015 came the first pure software-based, broadcast quality encoding and we soon also began to see content aware encoding.”
But as usual, with these solutions came problems. So the thrust of Diascorn’s presentation will be how AI can be invoked to help solve some of these challenges as we move forward.
“AI is very good at detection and processing large amounts of data,” he said, “and when it is empowered with machine learning it can be used effectively for decision making which can be more accurate than a human algorithm.”
When it comes to video compression, the bar for AI to surmount is maintaining video quality while reducing the amount of data required to create it.
After all, as he wrote in his paper, “Bit rate measurement is easy but picture quality is subjective.”
That’s why we’ll always need the human element at some point in the compression chain.
“Today all the video evaluation is done manually, but with AI coupled with Machine Learning, the improvements can be accomplished faster,” Diascorn said. “Because of that, we believe may new opportunities like customization will be opened for our industry.”
Jean-Louis Diascorn, senior product manager, Harmonic
During the presentation, he’ll be explaining that this is not going to remove the need for trained, professional intervention, because just like every improvement in technology over the centuries we will still need engineers who understand the goals of the communication industry and can give them meaning.
But codec algorithm design is not just about picture quality. As Diascorn will be discussing, since it impacts the density of the solution and the latency for live processing, we will again be looking to machine learning for improvements.
For example, by choosing the proper implementation and scope, the right machine learning algorithm can reduce the requirements placed on CPU processing.
The implementation of AI to advance the science, and art, of video compression is a topic that will impact all facets of the digital content creation industry, so it can potentially impact every attendee of NAB 2019.
So putting Monday’s “AI Technology is Changing the Future of Video Compression” session of the BEIT conference onto your game plan could be a valuable investment of your time at the conference.
After all, Jean-Louis Diascorn’s paper concludes with the hope that within the next 10 years we may get a new standard.
That’s something we should all be getting ready for.
Want to know more about this year's BEIT Conference? Click the link here to see the official schedule along with a snippet of information about each presentation.
Would a free exhibit pass help? Click this link or image below and enter the code MP01 at the correct prompt.
Here are some additional articles about Broadcast Engineering and Information Technology (BEIT) sessions taking place at NAB 2019 in which you may be interested.
You might also like...
Training neural networks is one of the most import aspects of ML, but what exactly do we mean by this?
As the wider broadcast industry picks up the pace with virtualized, cloud-native production systems we take a look at what audio vendors currently have available and what may be on the horizon.
One cannot get very far with electricity without the topic of batteries arising. Broadcasters in particular have become heavily dependent on batteries to power portable equipment such as cameras and lights.
The Sponsors Perspective: Proactively Monitor IP Video Networks & Essences With Inspect 2110 & PRISM
For over two decades Telestream has streamlined the ingest, production, and distribution of digital video and audio. Today, compared to its SDI/AES-based predecessors, IP video adds exciting new challenges to these workflows.
Information theory can also be applied to loudspeakers, which are among the most difficult of transducers to design. Measuring the information capacity of loudspeakers is a useful tool.