The conventional wisdom about cable TV is that you can’t beat the system.
Even if you’re subsisting solely on streaming video, or supplementing a cheaper cable package with Netflix and Amazon Prime, Internet providers are now threatening to make themselves whole again through data caps and overage fees. Comcast is now testing such caps in more than two dozen markets, and the industry as a whole is eager to follow, not because their connections are congested (they’re not), but because of the profit potential from people who must pay to exceed their monthly allotment. Barring some regulatory intervention, home Internet users will have to be mindful about how much video they stream, especially as data-hungry 4K video becomes more common.
Bleak as that sounds, what if data caps are a technical problem waiting to be solved? Just as smartphones have squeezed out better battery life through more efficient displays and processors–rather than just bigger batteries–perhaps improved video compression could save people from a nightmare scenario of steady overage charges.
Not only is this possible in theory, but quite a few groups are trying to pull it off. The only question is whether the industry as a whole can agree on solutions before home Internet data caps become reality.
In basic video compression, the trick is to figure out how much of the picture stays the same from one frame to the next. Instead of storing every pixel of every frame, compressed video tries to only store what changes.
The most popular compression format, called H.264, is more than a decade old, and ripe for replacement by something more efficient. Unfortunately, no one’s yet agreed on exactly how to do it.
A couple years ago, a successor called H.265, or High Efficiency Video Coding, emerged as a standard. But while it promises twice the efficiency of H.264, and is already used in 4K streams from Netflix and Amazon, adoption is not widespread in hardware such as set-top boxes and among streaming services.
That’s partly because the format is clouded by patent licensing issues. As Ars Technica reported in July, H.265 has two groups making royalty claims–MPEG-LA and HEVC Advance–the latter of which wants 0.5% of all revenue generated by streaming. By comparison, H.264 doesn’t draw any royalties from free streaming services like YouTube, and puts a cap on annual royalties for paid services. Given that H.265 requires new hardware–it’s not backward-compatible with the phone in your pocket or the 1080p TV in your bedroom–the industry is understandably skittish about jumping in. It’s a chicken-and-egg situation where streaming services avoid the new format, and so device makers avoid supporting it.
“When you don’t know what you’re going to pay, because there’s a bunch of uncertainty in the market, you’re going to delay actually putting that in hardware,” says Gabe Frost, executive director for the Alliance for Open Media, which is trying to create a royalty-free H.265 alternative. “Because once it’s in hardware, you’ve already bit the bullet.”
The Alliance for Open Media formed in September, with heavyweights Amazon, Google, Intel, Microsoft, Mozilla, Cisco, and Netflix as founding members. The group’s plan is to create something like H.265, but without royalties or patent headaches.
Frost says this is only possible by working backwards from how video formats and codecs (software that compresses and decompresses video) are typically created. Every member of the group has pledged to make their contributions royalty-free, and a big part of the work involves vetting those contributions to make sure they’re not at risk of violating existing patents. The Alliance will only get to building actual products once it’s sorted through these issues.
“The desire is to change the way these codecs are done, and deal with the [intellectual property] issues up front,” Frost says.
The idea of several tech titans joining forces to make a free compression scheme almost seems too Kumbaya to be true, but keep in mind the companies involved would all benefit even without royalties. For a firm like Microsoft or Google, which ferry heaps of data to and from the cloud, better compression would reap big savings.
“We see what’s happening today with the royalty uncertainty, and the reason why these groups of companies have come together is that they all have a business impact in not being able to move forward,” Frost says.
The Alliance’s biggest obstacle, then, might be convincing hardware makers to adopt whatever the group comes up with. The groups collecting royalties now include industry-leading hardware firms like Apple, Samsung, Sony, and LG, and they may not be as eager to adopt alternatives.
“At this stage, the Alliance of Open Media has issued only a declaration, is yet unproven in the market, and has not gained any momentum in media consumption devices,” says Frederick Savoye, vice president of the desktop business for RealNetworks, which is building its own proprietary format called RMHD. “The real challenge will come when the Alliance starts discussions with the industry leading device manufacturers, since most of them are members of the MPEG-LA video patent pool.”
Still, Frost says it’s important to offer a choice. “We think a next-generation codec from the Alliance answers this call for choice from the industry, and our membership–who each have made a royalty-free commitment–reflects that,” he says.
Even if the Alliance achieves its dream of royalty-free, high-quality streaming video, it’s going to take a while. The organization plans to have a specification ready in late 2016 or early 2017, and hopes hardware makers will adopt it in the 2017–2018 timeframe. In the meantime, several groups are trying to squeeze more bandwidth out of formats that already exist.
Beamr, for instance, says it can reduce bitrate of H.264 video by 20% to 40%, and it is already used by streaming services Crackle and M-Go. V-Nova works in a similar fashion, and has been offering its services to Internet providers who have their own video solutions. Both technologies perform an algorithmic sleight of hand, determining how to reduce bitrate in ways that users won’t perceive.
In talking to these companies, they’re happy to extoll the importance of standards and the work that the Alliance for Open Media is doing. But they also stress that they can move faster, and don’t require new investments in hardware and infrastructure.
“In reality, what we are selling at V-Nova is not an algorithm or royalties, we are actually selling working software, optimized to run on existing hardware in real time,” says Guido Meardi, CEO of V-Nova.
The good news is that compression needn’t be a zero-sum game. Even if the industry agrees on a next-generation format like H.265 (HEVC), or whatever the Alliance for Open Media is working on, there’s room for improvement beyond the format itself.
“The challenge is that there’s a fundamental constraint within the limits of the codec,” says Mark Donnigan, Beamr’s vice president of sales and market development. “Without innovation like what we have developed, which is actually outside of the encoder, outside of the codec, [H.264 has] really gone as far as it can go. The same trajectory is going to happen with HEVC.”
Even so, solutions like Beamr have their own roadblocks. Donnigan points out that streaming services would have to re-transcode their entire video libraries to adopt Beamr. Those services tend to store multiple versions of their videos at varying qualities, so they can switch between them based on the user’s bandwidth.
“You just start multiplying it out, and you realize that you’re talking 1.5, maybe 2 million hours of content that needs to be optimized, so it takes time,” Donnigan says.
It just takes time might well be the mantra of all compression efforts, but at least it’s a problem the industry is trying to solve. But will Internet providers wise up and reduce data limits further if compression improves? V-Nova’s Meardi suggests that won’t be necessary, as more efficient video will only increase people’s thirst for high-bandwidth applications.
“As soon as you lower the marginal cost, there are so many more things that you can do that have a value, that immediately demand skyrockets,” Meardi says.
Like a more fuel-efficient car that ends up putting on a lot more miles, the cruel twist is that better compression could only encourage more data use in the age of capped Internet. Beating the system might be trickier than we thought.