Grass Valley’s Sydney Lovely on the cloud, AI and the road ahead
Subscribe to NewscastStudio's newsletter for the latest in broadcast design and engineering delivered to your inbox.
Grass Valley’s 2022 NAB Show push is focused on transformation, showcasing how broadcasters can embrace the cloud and what it can enable.
With this in mind, we recently spoke with Sydney Lovely, Grass Valley’s CTO, about cloud production the continued expansion of the AMPP ecosystem and new technologies impacting broadcasters like AI and machine learning.
Our conversation has been condensed and edited for clarity.
Grass Valley has become more and more of a cloud services company. What is the breakdown in terms of where you’re finding most of your business now?
From my perspective, running R&D globally, about 70 to 80% of our engineering resources are software engineering and have been for quite some time. And even products like our iconic production switchers — they’re a complex beast with five million lines of code. So we’re not really new to the software game.
We’re still in reasonably early days in cloud adoption. And the way I would think about it is that technology falls into three main buckets. The first is the hardware-defined technology, like our production switcher is a good example of that, with complex software under the hood but people consume it through the form of hardware. The middle bucket would be products that are more software-based, and people have consumed in a traditional software way through on-prem servers or virtual machines. And those would be products like our playout systems or our MAM system, Stratus. And then the new third bucket is elastic or cloud computing capabilities. What we’ve seen is that over the last couple of years, those products that were traditional software solutions are moving pretty rapidly to the cloud, things like traditional playout systems.
Where are your customers looking to the cloud?
The things that are going to the cloud sooner are some of these derivative content or streaming channels, etc.
In the past, our customers had to stand up totally different solutions for their digital and streaming channels than they’ve had to do for their live production environment. As you can imagine, that’s a real hassle for them. So being able to do that in a more joined-up manner is really helpful to them.
In terms of building out the Grass Valley AMPP ecosystem, where are you in that journey, and what is the roadmap?
If you look at one of the biggest drivers of cloud computing economically — elasticity.
If I’m doing something only two hours a day, those types of things really shine in the cloud because you can spin it up, you pay for what you’re using and then you shut it down.
And the areas that are always on are less economically interesting. However, the juxtaposition here is that the technical difficulty of playout is by far the easiest and live production is by far the hardest. So that’s one of the reasons that this whole industry hasn’t really moved to the cloud faster than it has.
With our Grass Valley Media Universe approach, we said, “we have to solve the live problem first. And if we solve the live problem first, everything else should be downhill from there.”
We started with that live space and then just recently we’ve released an entirely new application suite around playout that lets customers deliver channels to air, brand them, subtitle them, monetize them with traffic, and so on.
We’ve also released our production application suite, with a native HTML editor, a fully elastic ingest service, etc.
We’ve really completed the core of the Grass Valley applications. Of course, we’ll continue to add applications to the ecosystem and enhance those applications, but at this point, you can build an entire TV station in the cloud based on the Grass Valley AMPP solution.
So now the shift towards opening up the platform to partners?
If you look at our customers, they’re under a lot of pressure due to digital disruption and the democratization of video in general.
They need things to be simpler and easier. So you used to have many, many, many, many different best-of-breed steps in creating a media supply chain, and each one of those things required specialized expertise on the part of the customer and the vendor community to pull all this stuff together.
We’re trying to really streamline this and make it far easier, so it really boils down to a simpler workflow for them.
Where do you see partners fitting into the AMPP ecosystem? Or are you looking at potential acquisitions?
At this point, it’s probably less about building it all ourselves or buying it. It’s probably more focused on the partner side.
We’ve spent five years building a true cloud-native microservices platform. And a lot of the partners we’ve talked to, there’s just no way they can get there on their own. But our customers need those types of solutions to be available in the cloud.
I think Andrew’s experience with building out those ecosystems is key, absolutely key. And so that’s something that has always been part of the vision for this. (Editor’s note: Lovely means Andrew Cross, Grass Valley’s CEO.)
Are today’s broadcast engineers comfortable with the cloud?
It’s a mixed bag. There are certainly traditionalists that are uncomfortable and if we contrast that with a lot of our more digital native customers, they’re extremely comfortable with it.
There are folks on the broadcast side that are still a little bit uncomfortable with it, but I’d say that’s really changed a lot in the last couple of years. Necessity is the mother of invention.
The platform we’ve built, it had an intention of making and giving those customers a familiar experience, so they don’t have to have a Ph.D. in AWS or in Kubernetes.
Would you say across the board that the industry is still lacking in the education necessary on this?
I think it’s changing. I’d say we’re beyond the first innings of the baseball game, so to speak. We’re kind of in the middle innings. So it’s a mixed bag, but some of our customers that came from very traditional backgrounds have learned this stuff very, very quickly. But there’s certainly a lot of demands on everybody because they need to keep their current stuff running. They need to adopt the new technologies and push forward.
How do you view new technology like AI and machine learning fitting into broadcast?
I see this unfolding in three phases. So the first phase is going to be AI and machine learning making what I’ll call editorial suggestions.
The next piece is going to be fully automated AI and ML with final director-level approval, we’ll call it. So it’d actually go through and create content, then you would approve it, and then eventually it’s lights-out, create it on your own.
Lights-out is created all by automation. So I think the lights-out stuff is a ways off.
We’re already there with the suggestive AI capabilities and even some of the capabilities that we’re able to take advantage of around creation with editorial approval, that second stage.
For example, our automated caption creation is integrated into our AMPP platform and you can use it either way.
Then I’d say on the live production side, we’re also starting to be able to see things like a capability around automated highlighting.
A lot of times with AI and ML, you think about it in terms of image analysis. It’s actually even simpler technology than that. It’s data wrangling. We’ve recently evaluated a partner technology that’s automated highlight generation where it’ll go through and take an entire soccer game, three hours long, and automatically generate a highlight reel.