Big Data Tackles the Multiscreen Video Quality Challenge

During a panel at the NAB Show in April, Hui Zhang, the CEO of the multiscreen video analytics firm Conviva noted one of the challenges of the streaming age: as the quality of video streaming over the Internet has steadily improved, our patience for buffering has diminished even more rapidly.

While video providers have made improvements in how they’re delivering streaming video over the Internet, customer’s expectations have raced ahead of these gains. Buffering may happen less frequently, but it’s far more alienating and liable to lose a viewer’s attention than it was even two years ago, Zhang observed.

And indeed, as Conviva has noted, only a small number of streaming video providers are actually delivering video at the highest bit-rate with minimal buffering. And that’s at home. On the road, on cellular networks, the experience is both worse and considerably more expensive as mobile network operators shed their unlimited data plans. To sustain the growth, customer satisfaction and ultimately the profitability of streaming services, quality control is a must.

Big Data to the Rescue

Streaming video is all about data: not simply the bits being sent over the Internet, but data about network conditions and end-point devices. So it’s no surprise that leveraging this information is becoming the critical ingredient in quality control.

Some video providers, notably Netflix and Google, have been collecting and publishing streaming data to call attention to Internet Service Providers about the quality of their last-mile connections. Armed with this information, consumers can presumably make a more informed decision about which ISP to choose (when they have a choice) if they’re avid streamers — or know where to pin the blame when buffering strikes.

Adaptive bitrate streaming, where video quality is dynamically regulated based on real-time data on network conditions, is another example of the power of data to regulate quality.

Finally, researchers at Bell Labs have provided a glimpse into a possible next-phase of data-enriched streaming, this time targeted specifically for mobile users. Dubbed “context-aware video streaming,” the technology “makes smart use of location-based information, including road and network coverage maps and radio performance data, to deliver algorithms that predict upcoming network performance. Based on this prediction, more seconds of the video are downloaded to the consumer’s device in areas of good coverage before the vehicle enters a poor coverage area,” according to a description released by Bell Labs.

Unlike adaptive bit rate technologies, which are reactive, Bell Labs’ context-aware solution is predictive, allowing it to anticipate and correct-for possible bottlenecks before they occur.

As Francis Bacon once noted, “knowledge is power.” Little did he know.