With the great success of Internet video and rapid growth of wireless data networks, mobile network operators (MNOs) and content producers (CPs) naturally are looking to mobile video as the next frontier and a future significant revenue generator. Unfortunately, mobile video offerings are currently hampered by the use of Internet technologies in the highly unique wireless environment. The result has been a lower quality of viewing experience (QoE) that could threaten the viability of mobile video.
 |
Dr. Sujit Dey
|
According to M:Metrics, former European mobile TV users outnumber current users by nearly 20%. Subscribers indicated dissatisfaction with quality, reliability and price.
Video quality is often viewed as a 1-dimensional issue of frame rate, based on the belief that "it's all about the smoothness of motion." It follows, then, that video QoE is seen primarily as a bandwidth allocation issue, just like on the Internet.
However, the very nature of wireless networks belies this 1-dimensional approach. Not only do wireless networks have significantly less bandwidth, but they are also intrinsically highly variable. Bandwidth availability can change dramatically and unpredictably from second to second. Increasing bandwidth can actually acerbate the issue, since the peaks and valleys are more severe. In addition, noise conditions not found in the wired world are an inherent part of wireless, resulting in packet drops and frame losses. Users moving in and out of coverage and between towers add to the unpredictability of the network.
As a result, there's fundamentally no such thing as wireless network Quality of Service (QoS). Unless network operators are able to drop or block users from services, they cannot divide up the available spectrum in a way that guarantees a specific application's QoS.
The correct measure to further user adoption - and retention - of mobile video services is the viewer's QoE. To measure QoE, we need to account for the network variability and noise conditions in addition to bandwidth constraints. And to do that we need to clearly understand the various quality factors:
Session quality: This top layer metric concerns itself with the user's overall experience in viewing the video. Initial buffering time - the time it takes the video to load and begin playing - and re-buffering during the playback are important factors in overall video experience. Synchronization between audio and video is also an important component
Video quality - Frames per second (fps) should not be ignored as an important metric. The human eye will in fact perceive a small difference between 20 fps and 10 fps on a small screen. However, the overall quality of each frame will contribute greatly to the user's perception of picture clarity. Moreover, excessive stalls will be perceived as "jerky," with a similar impact as a low frame rate.
Audio quality - Audio quality and fidelity also can affect perceptions of video quality. High-quality audio can overcome poor video quality, particularly for some content types, for example, a music video.
The nature of the content contributes to the weight of each factor in shaping QoE. A sports video may maintain a higher QoE in a stream that favors video smoothness over picture clarity. A "talking head" news clip may deliver a higher QoE by focusing on audio and picture clarity. It is important, therefore, that service delivery be flexible in modifying each of the quality factors to maximize the users' viewing experience.
As we've seen, mobile video quality is not a uni-dimensional problem solved by higher frame rates. A number of factors contribute to the user's perception of QoE, including initial latency, frame quality, stalling, re-buffering and audio quality. These factors are difficult to manage under the best of conditions, but are more acute in the highly dynamic, bandwidth-starved transmission path offered by wireless networks. While excess bandwidth, static bitrate connections, and the prevalence of PC clients may ameliorate these issues inherent to video on the wired Internet, such is not the case on wireless networks.
In order to maximize user QoE - and therefore mobile video adoption - video streaming systems must be architected for the low bandwidth, highly variable nature of wireless networks. The system must be able to shape the video stream in real time, across multiple video, audio and session quality factors, in response to each user's current network condition. Only then will Mobile Network operators realize the financial benefits of mobile video.
Dey is founder and chief technologist for Ortiva Wireless (www.ortivawireless.com).