In the heady early days of digital video, around 1995, video compression efficiency was at a premium. DirecTV, DISH Network, and other satellite operators worldwide were trying to pack more and more TV channels into a very limited number of transponders. The value of each transponder was so high that the market was willing to pay an amazing premium for any technical edge that allowed that bandwidth to be used more effectively.
There is a huge difference in compression efficiency between different video encoders, even if all those encoders are outputting video that is encoded to the same standard. Holding quality constant, one encoder might need 5 Mb/s to encode the same video that a better encoder can output at 3.5 Mb/s.
At the high end of today’s market, the difference between encoders is well understood, and companies like Echostar (the sister company of DISH Network which actually operates the uplink centers) still spend a ton of energy evaluating lots of different encoding options to get an optimal outcome.
Aside from the satellite players, though, most other video service providers have not tended to worry too much about compression efficiency. In particular, cable operators like Comcast and over-the-top (OTT) providers like YouTube and Netflix have lived in a world where bandwidth isn’t really all that expensive. As a result, they have placed less focus on super-efficient delivery of video down their pipes.
These providers tend to treat encoding technology as a commodity, and if one h.264 encoder is as good as the next, you might as well use one of the free open-source packages like x.264 or ffmpeg. And if minor problems exist in getting a fat enough pipe to the consumer, adaptive bit-rate algorithms such as HLS or MPEG DASH will seamlessly shift the consumer’s video quality to a lower level.
It’s likely that delivery efficiency is going to get more important in the future, though. A buzzword that has been gaining traction recently is “mobile-first” delivery of video. A while back, DirecTV made a hilarious ad starring the Manning brothers — your phone’s not for callin’, it’s for footballin’. Indeed, DirecTV’s NFL Sunday Ticket has seen so much success that the company just renewed its deal with the football league for an estimated $1.5 billion per year, and increasingly that video will flow to viewers over the Internet instead of satellite.
For their part, the U.S. mobile carriers have evolved their pricing to a “pay for what you use” model, so now they have a financial interest in creating more demand for data. Offering a better video streaming experience for customers is likely to be one key to finding new revenue. As evidence of this trend, Verizon recently announced a mobile-first video service that will debut this year and is expected to offer 20 to 30 channels.
So, for all these reasons, concerns around user experience and clogged bandwidth on mobile networks have started getting a lot more discussion.
To be sure, now isn’t the first time that demand for streaming video caused pundits to talk about the need for OTT compression efficiency. CNET reported in 1999 that “Despite advances in streaming media technology, the huge numbers of people unable to watch Wednesday’s live Victoria’s Secret fashion show Webcast showed the Internet still isn’t ready to handle mass market video.” Recent estimates show Netflix uses as much as 36 percent of downstream North American Internet traffic during peak hours, so we’ve clearly addressed some of the issues that plagued streaming video in its early days.
At the same time, there is only so much mobile bandwidth available. If you assume that delivery of video over mobile networks really takes off — which would be my bet — the entire industry is going to have to get serious about delivering this video as efficiently as possible. We don’t quite have all the incentives aligned yet: HBO Now doesn’t necessarily share Sprint’s interest in reducing data consumption. But winter, as they say, is coming.