State of the Union: Distribution Codecs
Video codecs have always been a subject of debate in our industry. With IBC2016 rapidly approaching, it’s the perfect time to refresh on the subject. Let me start out by saying that while HEVC has been deployed over the past several years, the scale of these deployments is being hampered by the licensing terms surrounding the new HEVC Advance patent pool. For Ultra HD distribution, HEVC is the only real choice. For SD and HD broadcast and OTT services, operators can choose to stay with AVC, although the bitrate will be twice that of HEVC. Where do we go from here?
There are always improvements that can be made to AVC, and Netflix has provided answers for VOD, as described in their technical blog. At NAB2016, the same technique was demonstrated for live applications. The main benefit of this approach is that no new decoder is needed; operators just need to upgrade the encoding infrastructure and add more computation. Today, that is an easy task thanks to the emergence of software- and cloud-based architectures.
As far as video codecs go, we see new initiatives cropping up all over the place. MPEG brewed up its Internet Video Codec (IVC), which was designed to be “royalty free.” IVC is a simple profile class of codec, and the performance will be lower than AVC. It’s expected to become a standard in 2017. Next on the codec map is NetVC, developed by IETF. There’s news that NetVC might merge with AOM’s AV-1, with standardization taking place by May 2017. Details can be found here.
AOM is a consortium of companies that wants to develop a royalty-free codec with twice more efficiency than HEVC in the 2017 timeframe. The Alliance for Open Media is governed by founding member companies including Amazon, ARM, Cisco, Google, Intel, Microsoft, Mozilla, Netflix and NVIDIA. Recently, Harmonic contributed to an article on AOM in Streaming Media magazine, describing the challenges AOM will face.
Last but not least, MPEG and ITU are working on the next-generation JVET (Joint Video Exploration Team) codec, aimed at delivering twice the bandwidth efficiency of HEVC. It will be standardized in 2019-2020. We at Harmonic believe that if new techniques such as elastic encoding, content-aware encoding , pre-post processing pairing and machine learning are applied in addition to a new codec, a quadruple improvement can be achieved for relevant applications. We are already participating in the JVET endeavors and, in due time, will demonstrate whether those claims made can be substantiated.
What is in store for the future? AVC evolution is not over, and we will continue to deliver improvements that make AVC comparable to HEVC. In addition, we see a lot of inertia toward moving to HEVC for legacy applications, not only due to licensing terms but also because the decoders have to be upgraded. In the OTT environment, where the client has a different dynamic (i.e., consumers purchase a connected TV or an IP STB) and there’s a short lifecycle for mobile devices of about 18 months, we anticipate there will be a lot more HEVC deployments. In the long term, all of the “magic codecs” making bold claims will have to demonstrate their performance, whether they’re actually royalty free and how they compare to more traditional approaches like MPEG, which have existed for the last 20 years. While newer codecs may deliver on the technical side, from a business aspect, there may be limitations, as has been shown recently with HEVC.
Harmonic, leading the video compression market with its acquisition of Thomson Video Networks, will always be at the forefront of video codec innovation and more importantly will always find a solution that fits the needs of its customers – both broadcast and OTT.
– Thierry Fautier, Vice President, Business Development & Innovation at Harmonic