Open Mic: Combining QoE and QoS for TVE services
TV Everywhere is an imperative for MSOs, but like all new services, it poses several challenges as operators strive to ensure quality while simultaneously expanding their multi-screen deployments.
Unlike traditional linear MPEG-2 services, TV Everywhere relies on the H.264 codec coupled with adaptive bit rate (ABR) streaming. Unsurprisingly, these new technologies pose quality concerns.
A comprehensive quality control (QC) process for TV Everywhere requires two stages in which errors in programming are first identified and fixed, and then the programming assets are correctly stored for seamless delivery.
Step 1: Content preparation – optimizing subscribers’ audio and video QoE The TV Everywhere delivery process starts at the headend where HD and SD programs are ingested for MSOs’ linear services. Errors that cause black screen, freezing and macro-blocking at the elementary stream level often occur in programs before they reach the headend. This makes identifying and correcting them at ingest the most important step for ensuring an end-to-end error free video delivery.
The next step of the process is to transcode these programs into multiple representations of H.264 video and AAC audio for different viewing devices, including smartphones, tablets and computers. Transcoding often creates compression artifacts and other picture and sound quality errors that degrade subscribers’ quality of experience (QoE). Many MSOs use advanced, real-time monitoring systems to identify, alert and capture these problems. Some of these monitors are capable of maintaining 60-day historical reports to aid troubleshooting and forensic analysis and can also detect IDR (Instantaneous Decoder Refresh) and close captioning issues, which may degrade subscribers’ QoE, cause non-compliance and lead to ad insertion failure.
Ensuring that content is error-free after transcoding is vital because the next steps are for the programs to be fragmented, packaged and encrypted, at which point the large volume of video and audio content makes QoE monitoring not cost effective. MSOs must now turn to quality of service (QoS) to ensure that their pristine content is streamed seamlessly to subscribers when requested.
Fundamentals of ABR ABR streaming is fundamental to TV Everywhere as it addresses the bandwidth fluctuations inherent with the network and optimizes user experiences across different viewing devices.
ABR addresses this by chunking a single program into three-10 second “fragments” at multiple transmission rates and profiles. These options are called representations. End-user devices request the most appropriate representation based on the device type and the currently available bandwidth. As bandwidth fluctuates, the service “toggles” in real-time among different representations so viewers always receive the highest possible quality stream for the available bandwidth and do not experience any pauses.
Step 2: Content delivery - optimizing network QoE for TV Everywhere At this point, with the programming encrypted and stored in the origin server in different representations as fragments, monitoring QoS on video asset availability and network performance becomes vital.
For example, when a tablet requests a fragment the system must check the manifest file and ensure that the correct bit rate is indeed sent. If the tablet requests fragments at 500 Kbps but is incorrectly sent fragments at 1 Mbps the service may stall.
Another reason QoS is so vital is that ABR streaming uses the Hyper Text Transfer Protocol (HTTP), which is a unicast approach in which only the stream requested is sent. This puts tremendous pressure on the system to deliver data precisely. ABR’s complexity requires a rigid and structured test and measurement regimen to ensure the system functions correctly.
Monitoring and validation are accomplished in “active” and “passive” modes in the QoS environment. Active devices can audit the asset, server and network integrity by proactively initiating multiple HTTP sessions as if multiple users are making requests. This is a vital QC step during trials and for “bulletproofing” large-scale deployments as it can help operators uncover asset, network and server issues during early roll out and trial periods.
Passive devices monitor QoS of the end-users’ initiated sessions. Passive monitoring is particularly ideal for statistical QoS analysis based on users’ actual sessions.
Summing it up To summarize, operators must employ comprehensive QoE and QoS monitoring to ensure the best possible user experience for their TV Everywhere services.