By Sean McCarthy, Ph.D., Fellow
A decade ago when Advanced Video Coding (AVC/H.264/MPEG4 part 10) was first introduced as a new international compression standard, it was hailed as improving video quality and reducing bit rates by a factor of 2 compared to MPEG-2 video compression. Today, the buzz is about a newer new international compression standard, this time called High Efficiency Video Coding or HEVC for short.
It is tempting to repeat the old theme that HEVC is twice as efficient as AVC just as AVC was twice as efficient as MPEG-2. But doing so would ignore history.
AVC did much more than double compression efficiency. It enabled new distribution models such as IPTV. And it changed the way we watch television by enabling tablets, smartphones, and laptops to become mobile TVs.
HEVC, the newcomer, could now change the heart of cable television’s infrastructure. In doing so, it could finally make the consumer’s dream of watching whatever one wants, wherever one wants, whenever one wants, into a bona fide everyday expectation.
The digital television industry has been moving towards the whatever-whenever-wherever model for awhile. We started by putting personal digital video recorders (DVR) into set top boxes. We also created an infrastructure and back office that lets consumers order programming on-demand over cable pipes. More recently, we’ve been leveraging IP networks and transcoding to deliver programming to both traditional and new mobile television screens such as tablets and laptops.
Logically, the next step is to put the DVR functionality in the network (nDVR or cloud DVR). Yet we have been largely shut out of that option because the amount of storage and file management needed have been too monumental. That is where HEVC comes in, but not in the most simplistic way one might imagine.
Just as AVC needed IP-encapsulation as a partner to enable IPTV; HEVC needs Just-in-Time (JIT) transcoding and JIT packaging as partners to give viewers nearly unlimited viewing options.
A whatever-whenever-wherever nDVR can be created by brute force. Save a copy of the MPEG-2 program indicated by the subscriber. Then create bouquets of AVC/MPEG4 and HEVC profiles to support multiple screens and bandwidth options. Next, individually package each set of profiles for delivery over HLS, Smooth Streaming, MPEG DASH, etc. Finally, install a file management system to keep track of this explosion of content fragments.
Brute force is not always the simplest solution.
A better alternative is to store a single “reference” copy of the subscriber’s content. Install a just-in-time (JIT) transcoder to convert the reference content to the bit rate and compression format needed to support the subscriber’s viewing request. For traditional cable services, the reference copy can be transcoded to MPEG-2 and sent by QAM. For multiscreen and adaptive bit rate services, the JIT transcoder can create the individual profile variants on-the-fly as needed and deliver them to a JIT packager to create the appropriate HLS/Smooth-Streaming/MPEG-DASH formats.
The JIT transcoder/JIT packager architecture can cut storage by at least 75% because each set of AVC profiles can easily grow to be as large as and larger than original MPEG-2 program the subscriber wanted to record, and one would need to plan on creating three adaptive streaming variations. With JIT transcoding, the AVC profiles do not need to be stored. With JIT packaging, the adaptive bit rate variations do not need to be stored. Only the “reference” copy takes up space.
HEVC will always be introduced as the smartest and most efficient member of the MPEG compression family for the foreseeable future. But deeds are more important than promises. We think HEVC will be remembered not only for its innate compression efficiency, but also as the cornerstone of the next-generation cable television architecture that finally created what consumers have always really wanted: The convenience of having a high quality television experience wherever and whenever they choose.
Background Notes on viewing behavior and just-in-time packaging:
1) ARRIS 2014 Consumer Entertainment Index (http://success.arrisi.com/arriscei)
2) L. Milin & C. Ansley. (ARRIS) “Cloud-Based DVR and Multiscreen Support Strategies – Optimizing Storage and Transcoding.” The Cable Show, Spring Technical Forum, 2014 (http://www.nctatechnicalpapers.com/Paper/2014/2014-cloud-based-dvr-and-multiscreen-support-strategies-optimizing-storage-and-transcoding)
3) “REPACKAGING Overview” (ARRIS/SeaWell Network) (Spectrum_Repackaging_Overview-1.doc)