Making the quality of online video matter
4th April 2017
There is no need to argue about the benefits of adaptive bitrate (ABR) as a main method of streaming video on mobile devices – it makes content accessible to viewers in a more efficient way, eliminating buffering, adjusting to the available bandwidth with as little impact on picture quality as possible. As ABR streaming becomes more prevalent, the need arises for standardization and better understanding of video quality of OTT services.
As the leading industry organization in the ICT space, the ITU (the International Telecommunication Union) launched a standardization procedure for the development of methodology for estimation and evaluation of the quality of ABR streaming. Ericsson, together with six other companies, contributed with recommendations for the development of video quality estimation models. The contributors included representatives from the telecom sector, other network technology vendors and research companies. Following ITU’s evaluation of the submitted recommendations, the companies worked collaboratively to merge their model proposals into a single model to deliver the most accurate quality estimation. The final model was approved as ITU-T Recommendation P.1203 in December 2016.
What is the model about?
In simple terms, the ABR quality estimation model is a formula for the calculation of perceived video streaming quality of service using objective parameters. Some of the parameters include resolution and framerate, A/V encoding rates, packet-loss rate, rebuffering and video freezing events. It is an integrated model that not only combines video and audio quality into the estimation of a single event, but applies a longer term evaluation to get an assessment of the user’s perception while watching a longer video. The data necessary for the estimation is collected mostly from the video player and the nodes or probes in the delivery network and the outcome is available for the media company to analyze.
The quality estimation model is just Phase 1 of an ongoing standardization process for streaming video. In the next step, ITU will work on a model that can be used for estimating quality of 4K, UHD content encoded by H.265/HEVC (High Efficiency Video Coding) and VP9.
You can learn more about the recommendations and the estimation model in my colleague Gunnar Heikkilä’s blog post here.
Transparency, control, experience
Media companies have been actively using adaptive bit rate technology for many years and it has essentially become the standard practice for high-end streaming providers allowing little buffering when streaming high-resolution feeds. The importance of the quality estimation model for OTT and TVE service providers is significant. It improves transparency and understanding of the quality of service. A standardized way of estimating quality, allows for an easier comparison of the services of OTT operators, which in turn leads to better competition and ultimately better services for the viewers.
If the technology is working properly, the content should playback without interruption. Even when users cannot notice any degradations, the quality estimation technique can detect small declines in quality when they occur. This allows OTT providers to improve the services promptly before the problem impacts viewer experience and satisfaction. Having a better understanding of the quality of service means more control and proactively solving issues in the platform. It also brings efficiencies as it can direct investments to the online video platform features that directly impact quality.
The quality estimation model is already being put into use in Ericsson managed OTT services. The data from the estimation will be visible in the analytics and reporting modules. It will provide customers with additional insights necessary for building the desired viewer experience and engagement.
Stan Dimitrov, Product Marketing Manager, Online Video Services and Sports Graphics