• Home/ Blogs / Mile High Video Preview: Optimizing Real-Time Video Encoders with Machine Learning

Mile High Video Preview: Optimizing Real-Time Video Encoders with Machine Learning

By Nelson Francisco, Principal Video Compression Engineer, MediaKind

This week, I’ll be heading over to Denver for ACM Mile High Video (MHV), a leading video coding and streaming event at the Colorado Convention Center. MHV is among the first major events in our industry after the COVID-19 pandemic, and it’s certainly an exciting prospect to be back on the showfloor after many months of video calls.

The media delivery industry has undergone a deep transformation during this period, adapting to assume a fundamental role in people’s life when they were unable to socialize or do many other activities. Media technology has developed in tandem in incredibly exciting ways. That’s why I’m excited to be sharing the MHV stage with my colleague Julien Le Tanou to present a session on “Optimizing Real-Time Video Encoders with Machine Learning (ML)”. We’ll be taking to the stage at 11:40am on Wednesday, March 2.

The significance of real-time video encoding

MediaKind is constantly looking for methods to improve the bandwidth efficiency and energy footprint of our encoders and provide state-of-the-art products and technologies to our customers. Throughout the past few years, we have investigated new disruptive algorithms and techniques that can bring added value to our portfolio. Artificial intelligence (AI) technology is just one example that has experienced accelerated growth in popularity over the past few years, demonstrating why it holds huge potential and versatility. AI is already being widely adopted in applications ranging from autonomous driving to natural language processing.

AI’s benefits also extend into image and video applications. However, the nature of current research is not always completely aligned with the problems and challenges players in the media delivery industry face today. AI has allowed us to analyze data trends faster and more efficiently using accelerated algorithmic design and specification. It has proven an extremely effective method for automatic detection of anomalies and behavioral issues within media workflows, by detecting potential problems earlier in the development process. From a product functionality perspective, our AI-based compression technology has enabled us to improve bandwidth efficiency while reducing resource consumption.

Join us at MHV 22!

Julien and I will be focusing our presentation on the latest developments and applications of AI and ML for media, including improving the video quality on real-time encoders with limited computing resources. We will talk through how MediaKind has developed new approaches to encoding which use low complexity encoders that adapt to both high-level and low-level encoding strategies. This can either maximize the video quality or further reduce computational requirements, bringing significant savings on hardware and energy costs, with clear environmental benefits.

I am delighted to be attending MHV 2022 and to be part of a rich line-up of technical presentations and panels. Some of the most prominent players in the media delivery industry will be in Denver, and it’s a fantastic opportunity to learn more about the latest trends and evolutions impacting the codec landscape. Stay tuned for a future blog, where I will be reporting back on some of the new ideas, concepts, products, and technologies from the event that I believe will play a big role in shaping the future of our industry.

You can find the full agenda for the event here.

If you’re attending MHV 2022, please feel free to connect with myself or Julien via LinkedIn, and we would be happy to meet with you at the event!