Broadcast to mobile
One U.S. mobile operator, MetroPCS, which has almost 10 million subscribers, has come down in the broadcast camp, after announcing this past January that it would sell an Android-based Samsung smartphone later in 2012. The device will be equipped with an ATSC chip, the hardware needed to receive broadcast mobile television signals in the U.S. The phone will be preloaded with the Dyle mobile television app from Mobile Content Venture (MCV) — a group of 12 broadcasters that came together in 2011 to provide a nationwide mobile TV service. But, Springall does not believe such services will gain much traction, not least because users are unlikely to watch much long-form content on mobiles.
“On-demand works well for mobile, content snacking and so on, while live or ‘event’ television is more social,” Springall said. “The big game is better enjoyed on a big screen with your friends.”
Springall’s view is that sending chunks of video via some form of HTTP streaming as in MPEG DASH will work well enough for delivering the limited amount of broadcast content that will be required.
That view, however, is not shared by everybody in the mobile video world, given the proliferation of tablets that are better suited to live TV with their larger screens than mobile phones, as well as growing video consumption on laptop and netbook computers.
But, nearly all video consumed on these devices is delivered over Wi-Fi at present, given that it is installed in most homes, generally having more bandwidth and providing better quality than 3G and often than LTE 4G. But, the quality in both cases depends on the number of users sharing a given cell, and this comes back to the broadcast argument. Within a home Wi-Fi network, there is usually enough capacity to support one or two users consuming video on laptops or tablets. Within a public hot spot, however — at an airport, for example — there may be a number of people all watching the same popular channel. This can bring the network to its knees, resulting in slow response for other users, as well as diminished video quality for those watching TV.
An experiment by the University of Wisconsin-Madison found that as few as three users attempting to watch an HD video stream over Wi-Fi resulted in a sharp drop in performance and quality. This is partly because the video packets are sent separately in unicast mode to each of the three users even though content is the same. But, the study identified another factor, which is that the Wi-Fi network places equal priority on all packets without knowing how relevant they are for the application. This is inefficient because, for video, some packets can be dropped with much slighter effect on quality than others. This led UW-Madison to develop software called Medusa, which implements video broadcast over Wi-Fi. It incorporates priority mechanisms that place greater effort on delivering important packets, such as packets carrying bits of the I frames within compressed video that contain information needed to display subsequent frames. Medusa has been demonstrated delivering HD video consistently at 20Mb/s over Wi-Fi — more than enough for full broadcast-quality.
It is quite likely, then, that broadcast techniques such as Medusa will be deployed within Wi-Fi hot spots that entertain many users, even if not for 3G or 4G cellular services. The ability to cut in and start broadcasting as soon as two or more people are consuming the same video would have a large impact on quality of service, not just for the people watching TV but for other users on the network as well.
It may well be in any case that more video is consumed over public Wi-Fi than 4G, given the widespread availability of hot spots in places where people spend their time while on the move — transportation terminals, restaurants, etc. In that case, it is quite likely that mobile broadcast will at last become an important medium, even if not over cellular networks as was originally intended.
—Philip Hunter writes the “Beyond the Headlines Europe” e-newsletter.