There’s a massive shift coming in the professional video equipment markets. Sure, we’ve seen shifts before: black and white to color , standard definition to high definition , analog to digital , etc. But I think what’s coming is more fundamental and the likes of which our industry hasn’t seen before: commoditization .
Introduction
The professional video industry is an industry at all by the fact that, technologically speaking, it needed to be pretty cutting edge. Moving video, first in analog and then digitally, was a high bandwidth affair. As such, required an impressive amount of hardware designed for purpose. I witnessed this first hand in my video disk recording business a decade and a half ago. Working with HD and 2K material required a hefty chunk of hardware to get it done. Sure, the industry has progressed towards more cutting edge tech since then: 3D, 4K, HDR, 8K.. but with each step two things happened: there was less and less consumer adoption and the particular video technology continued to fall behind state-of-the-art with regards to the more general tech industry. This meant each step was not only less lucrative (since fewer adopted it), but also each step required comparatively less hardware and expense.
And now we find ourselves where we are today. Sure, there’s lots of hard problems to solve still. But the technology backing those problems isn’t, practically speaking, bleeding edge. Further, the general compute industry has made massive gains: solid state storage is incredibly fast. What I was doing with 24 spinning rust arrays only 15 years ago now gets done by a small little M.2 board attached with a few screws. Networking has come down in price massively: 10 gigabit Ethernet is now easily affordable, with 25 gigabit on its heels. And GPUs, holy crap. They’re amazing pixel processing machines capable of being programmed in general purpose ways (which of course made possible the rise of AI, but that’s a different story).
Timing.. isn’t everything it seems.
I’ve spent the last few years working on SMPTE 2110 products. I’d think everyone in our industry is aware of ST2110 these days, but if you aren’t the TL;DR is: it replaces SDI and SDI routers with Ethernet and Ethernet switches. Common, off the shelf switches. Sure, they’re expensive. But not million dollar plus SDI router expensive. And they’re repurposable. They can move ST2110, but just as easy move NDI, or SRT, or H.26whatever. It’s just data over a network.
And yes, I hear you. But we need low latency and low jitter for pro video. ST2110 does a lot to ensure we meet some tight timing constraints. To which I say, sure. But not even close to as tight as SDI. And, if we’re talking ST2110 wide, then it’s barely tighter than an internet video stream.
But this isn’t about ST2110. I think ST2110 is a stop along the way to something much bigger. Once these massive data networks replace SDI infrastructure, then it’s a race towards commodity. I think the near future is asynchronous video running on nodes of interconnected commodity hardware. We get to a point where timing still matters, but it’s not the same constraint it has always been. We don’t need things to be genlocked, just locked. We’re not going to be mixing pixels in an FPGA, we’re going to be processing lines or blocks of lines in a GPU. That got its data from an off the shelf network board capable of RDMA . And yea, PTP solves the lock. Heck, there’s even a standard for using PTP to discipline the PCIe bus itself (called PTM ). And we’re going to be managing end-to-end latency as a complete network problem instead of worrying about the “frame delay” of a particular piece of equipment.
What to do?
So where does this leave our industry? Well, to be frank: in a software defined world. Specialized video hardware will simply become unable to complete on a price + performance + flexibility standpoint. My fears are this opens opportunities for industry players to build walled gardens, basically breaking the standardized inter-op that our industry enjoys (and is frankly one of the reasons I love it so much). While certainly profitable for those companies who do it, at least in the short term, it’ll be a massive negative change to our collective culture.
So, what to do about this unstoppable march towards commodity and closed platforms? We can start working to standardize the base level of technology and software. Ideally, we’d end up with a few open-source base libraries in which every company could then build their best of class video thingy on top of. We can push back against companies who are flirting with vendor lock-in and urge them to play nicely and openly. And finally, we can start learning all about this networking and IT stuff because it’s coming and there’s no stopping it.