IBC 2019: Codecs Top the Agenda
After years of being consigned to niche corners of the industry, codecs are now all the rage. In part that's because the need to cost-efficiently deliver increasingly high bitrate content is outpacing standards-led codec development, and in part it's because of the gap left in the market by the collective failure of HEVC patent holders to offer predictable licence costs.
At IBC, beginning in Amsterdam later this week, the debate busts out into the open.
V-Nova, for example, will be promoting the merits of LCEVC—or MPEG-5 Part 2—whilst carefully avoiding mention that the Low Complexity Enhancement Video Codec is based on its own Perseus Pro.
MPEG's endorsement of Perseus Pro underlines the claims made by V-Nova that the addition of the technology to codecs already in the field— including H.264, HEVC, and AV1—will improve video quality while reducing processing power.
While LCEVC has yet to be ratified by MPEG as a standard, it is no surprise that the first implementation of it is in the Perseus SDK, which V-Nova has now renamed P+.
Among a number of contenders seeking airtime for their compression tech in Amsterdam is startup iSize. Like V-Nova, it is applying machine learning to finesse compression, and again like V-Nova, it claims to offer a codec-agnostic performance boosting lightweight implementation.
iSize is coming to market late in the day but does so with the benefit of patents based on deep neural networks.
"Instead of internally customising encoder options in a static manner or using boilerplate streaming recipes that are not tailored to the input content and encoder, [we] offer the new concept of deep video precoding (DVP)," co-founder and CTO Dr. Yiannis Andreopoulos will explain in an IBC technical paper.
The core of the codec debate will take place on the conference floor. Real-time decoding and AR playback of MPEG's video-based point cloud compression standard is being presented by Nokia's Sebastian Schwarz; while Jonatan Samuelsson of Divideon argues for the emerging MPEG-5 EVC standard, which is effectively a royalty free version of HEVC.
Notwithstanding AV1 being under a cloud of its own (now that patent pool holders including JVC, Toshiba, and Orange have tasked Sisvel International to protect their interests), expect the AV1 ecosystem to be fleshed out with new product at the show.
In a whole day of sessions sponsored by Intel (to market its Visual Cloud processing firepower), speakers from Netflix, Google, Facebook, and Tencent will talk up AV1 for VOD and live streaming.
They can be expected to focus attention on Scalable Video Technology for AV1, a version already announced by Intel and Netflix and designed for encoders to scale their performance levels based on the quality and latency requirements of target applications. SVT-AV1 is capable of running a 4K video stream at 60 frames per second at 10Mbps (using the Intel Xeon processors, of course).
It's also open source for AV1 developers like Harmonic, Bitmovin, and Beamr to lend support.
Automation Boosts Live Production
A rack of new product, or rather evolutions of existing ones, promising to automate production with machine learning will dominate the live production space. AI is marketed as the silver bullet required to create, manage, distribute more content to more people with less budget.
The narrative is that AI-infused ingest to archive systems can master the complexity of live production so that broadcasters can maximise their creativity by focusing on story rather than technology.
This will make more sense once entire workflows are ported to the cloud and video is automatically enriched with metadata tagging,facial and object recognition and speech to text.
Among vendors with new product in this area are MAM developer Tedial with its sports-centric platform Smartlive; LiveU, which is marketing a live sports data-based production platform with "smart race car" manufacturer Griip; Dalet, which is arguably the MAM leader in enterprise applications for global news organisations; and IBM, which is sponsoring IBC's behind-closed-doors leadership forum and will present the automated sports clipping prowess of its AI engine Watson.
Perhaps the most eye-catching product, not least because it is already being used, is TVU Networks' cloud-based solution for multi-camera production that enables remote audience participation with video.
"The ability to combine metadata generation with timecode information further speeds and simplifies the search process, enabling producers and the AI to locate the exact content down to the video frame," says David Jorba, EVP, TVU Networks. "From there producers can customise content production tailored to individual audience preferences and platforms."
Sky Sports' live chat show, The Football Social, uses TVU Talkshow, as has new L.A.-based streaming channel Fox Noire.
Data Drives Personalised Experiences
"The biggest challenge the industry faces in the drive to optimise the user experience is understanding available data, predicting viewers' motivations, and ultimately using this data to create viewer-tailored experiences," says Peter Szabo, department lead UX/UI at solutions provider 3SS.
Machine Learning has become one of the most crucial research topics in UX. If users can be categorised based on their behaviours, interfaces can be tailored to their needs, ultimately creating the ideal, custom journey for each viewer.
"This categorisation problem is now solvable, thanks to the research related to deep neural networks," says Szabo. 3SS come to IBC with a mathematical formula to describe this television experience.
IBC 2019 kicks off on 13 September in Amsterdam, and here are some of the key things to see and do when it comes to the world of streaming media.