Broadcast in the Cloud

 A very prominent topic of this year was the migration of on premises infrastructure to the cloud. As a concept, this has been around for a while but it felt like there was a real push to help the industry really understand the benefits and challenges to this approach. I sat in a number of talks and discussions around this topic and the overriding hurdles from a broadcast perspective is the concern for copyright security of sensitive content (especially pre-release) and the sheer size and quantity of digital content these cloud workflows would need to support, in terms of moving files around flexibly. There was as much focus on the workflow point of view as there was with the distribution to the consumer aspects from these ‘new-look’ infrastructures. Interestingly, there wasn’t too much discussion around ‘vendor lock-in’ with the cloud which surprised me, given the limited ‘big players’ dominating this area. Many of the talks tended to be from intermediary companies whose focus was to help content producers manage their workflows and content within the cloud. Make.tv suggested that a good approach to getting into the cloud was to initially use it to ‘fail fast’ alongside current legacy on premises infrastructure. This helps smooth the transition and makes sense. They also spoke about how ESports is interesting from a broadcast perspective as it requires extremely low latency and therefore very low tolerance from users. As a result, workflows for ESports must operate as close to the edge/user as possible. Broadcast can learn a huge amount from ESports and the presentation drew attention to the fact that worldwide audiences for some events are bigger than major broadcast events such as the Superbowl. ESports also had a large focus at this years’ IBC including a final day ESports showcase (however I unfortunately didn’t manage to get to it).

Shift to Open Source (& the Cloud)

Intel are looking to impact heavily in this area by helping to develop open source services in their ‘Visual Cloud’ project. This aims at allowing companies to innovate  quickly using these tools and essentially hoist the optimisation effort up to the community rather than, for example start-ups having to spend resources on these types of tasks and instead be able to focus on the innovation itself. The Intel Visual Cloud covered many areas from gaming, media processing, analytics and more, which is a huge but interesting remit for Intel to want to take on. I think this type of approach is also interesting whereby a traditionally hardware focussed company like Intel are looking more and more like a software company as they adapt to the changing landscape. It was also intriguing that they don’t appear to be going head to head with the likes of Amazon and Google in physically providing the infrastructure but it was unclear which platforms the Visual Cloud projects would run on. Potentially it is platform agnostic, which I think would be the smartest approach to take. Open source was also a topic that seemed to appear a lot this year. Several talks were very positive about embracing the use of open source software and as a Software Engineer, it was great to hear large companies seriously looking to contribute their efforts in this area as opposed to developing their own closed solutions. I feel that this will increase innovation and development speed of technology, with the benefits for the large companies being that some of the responsibility of development is abstracted to the community. I saw a talk from a company called ‘Collabora’ who help companies utilise the power of open source software throughout the broadcast chain to build their own solutions. This is a great idea and something that maybe 10 or 15 years ago might have been a much less viable business model. I expect that there will be many more of these around as the open source software approach within broadcast grows. Hollywood’s vision of the future (2030) continued the topic of the cloud with asset storage and workflows all being stored there. However there were concerns that security and access management had to satisfy confidence before the big high value productions could entertain this. As a technologist, I felt a little underwhelmed by this vision. I feel that by 2030 Hollywood should be well beyond whether they should be using the cloud or not given the impact the cloud is already having today in broadcasting. I expected Hollywood to have had more blue sky thinking on their future workflows, especially given the panel was made up of representatives from the likes of Disney, Sony, Warner and Universal. An interesting thought that was put forward, however, was the desire for a common API for cloud based production workflows. The panel said that this would be important so that it could allow content creators to choose their own tools but remain compatible with each other over the cloud. The huge bonus of this is that then talent could be utilised from across the globe, resulting in more diverse working, creativity and reduction in production costs as people do not need to be on location.

Individual AI Innovation

Individual AI based innovation projects also seemed to be a topic with strong presence, with different companies and broadcasters keen to demonstrate what they have been working on and then looking to drum up collaboration with companies and broadcasters keen to help continue  development. It was good to see that a collaborative environment was being sought to develop the best solution and not multiple solutions that all do the same job but no better than each other (which tends to be common within Broadcast technology). The future influence of AI included talk about ‘V-Tubers’ which was virtual characters (or bots) creating content. The discussion around the issues of deep-fakes also came up off the back of this which is a major concern generally but it drifted a little from broadcasting specifically.

Personalised & Immersive TV Experiences

I watched an interesting discussion on mood states and the use of data to personalise experiences. Thoughts here centralised around the idea that listening to music is very mood influenced and in a way, ‘mood management’ and so maybe tapping into a person’s mood might be another key dimension to developing effective personalised recommendations for broadcast content. A big issue here is that this isn’t being done in TV yet, mainly because of lack of skills, knowledge and data science resource. Getting the balance right about how much you know about a consumer is important ethically. Personally I think the innovative applications for AI in broadcasting are exciting and I am also encouraged that the question of ethics is a central part of this too.

5G for Broadcast

5G was also a prominent topic as expected. The rollout of 5G will be critical to changing the way the broadcast industry evolves. There was an interesting presentation on a paper from an IEEE member around designing 5G networks for future media delivery. The takeaways from this were how we should try not to build 5G networks in the same way we build LTE/4G networks as we’ll potentially run into similar limitations. Instead, they suggested a ‘one to many’ multicast/broadcast approach with dynamic switching to unicast. This would ensure that these networks could be better suited to use cases such as VR which will feature heavily in the broadcast industry’s future, whilst also harnessing benefits like less network load inherent in multicast and broadcast networks. The opportunities here sound exciting if the right decisions on this are made at the right time. Game of Thrones, The Lion King and Toy Story 4 on the 8k Dolby Atmos Auditorium screen was also a great set of experiences, particularly the sound from the 10 point Dolby Atmos setup. During the Game of Thrones talk beforehand, it was also interesting to hear how HBO did things differently to the norm, such as shooting entirely digitally (rather than on film) which saved around one million in the first series alone.