Looking back: hot topics at IBC 2017

As the 2017 edition of one of the biggest trade shows in the broadcasting business has just finished, this blog post provides an overview of interesting topics presented at this year's International Broadcasting Conference in Amsterdam.

Microsoft embraces CBC encryption

When CMAF was announced last year, its promise to deliver a truly unified media format was received with some skepticism as it became clear that two incompatible modes of encryption remained part of its specification. 

CMAF will not only be a unified format for content in the clear, but for encrypted media as well.

Even with CMAF, it looked like serving encrypted content to a wide audience would necessitate encrypting the content twice to cover the major DRM systems: using Cipher Block Chain (CBC) mode for Apple's FairPlay and using Counter Mode (CTR) for Google's Widevine and Microsoft's PlayReady.

This changed somewhat earlier this year, when Google announced CBC support for Widevine. With Microsoft's now also integrating CBC support into their products with PlayReady 4.0, CMAF will not only be a unified format for content in the clear, but for encrypted media as well.

Cedexis integrates Mux's video performance analytics

The internet is a busy place and, for companies offering content on a large scale, multi-CDN strategies have become the norm: if one Content Distribution Network (CDN) experiences problems, a multi-CDN strategy ensures that there is another one that can be relied on to get content to clients as quickly as possible.

Cedexis' technology has become a vital part of such strategies, as the company offer services that track the status of internet delivery through 14 billion real user measurements every day and then route traffic around congestion and outages, using the CDN that offers the best performance at a given time.

Mux collects and analyzes client-side streaming video events that define a user's quality of experience.

This high-level approach works well for almost all traffic, but offers room for improvement when it comes to video. This is where Mux fits in, as it collects, processes, and analyzes streaming video events from the consumer's video player to quickly identify the events that define the quality of experience.

Now that Cedexis has integrated Mux's video performance and quality analytics into their services, clients of theirs who are in the business of delivering video will be able to guarantee customers an even higher quality of service.

Wowza ClearCaster hardware (Facebook Live)

In a surprising turn of events, software vendor Wowza launched its first ever hardware product at IBC. Developed in cooperation with Facebook, the sole focus of the rack mountable ClearCaster is professional grade streaming to Facebook Live.

Catered to Facebook Live's specific feature set, ClearCaster is about more than getting a high-quality stream up and running quickly, as it integrates Facebook's 'talent view' on the production side. This view presents a content producer not only with the image that is currently being streamed, but, thanks to Facebook Live's API, overlays it with live responses from viewers.

Without a doubt, the ClearCaster is catered towards industry professionals, with the incorporation of an SDI input and a significant price tag to boot. Currently, it supports Live streaming of 1080p content at 30fps, with encoding based on the open-source x264 component.

Hybrik's cloud-native Live encoding

At this year's NAB in Las Vegas, Hybrik was awarded a Best of Show award by Streaming Media for its Amazon AWS-based cloud-native encoding offering that has made its mark with highly competitive pricing.

The setup uses Unified Origin for a cost-effective setup that can serve out content to all devices.

Until recently, Hybrik only offered on-demand solutions, but at IBC they demoed a livestream as well. Their setup made use of an Amazon instance running Unified Origin for a complete and cost-effective setup that can serve out content to all devices.

CTA's WAVE project

In an effort to further enhance the quality of internet-delivered commercial video, the Consumer Technology Association started the WAVE project, which builds upon existing standards such as MPEG-DASH, Apple's HLS, HTML5, EME, MSE and CMAF to define the requirements that are necessary to close the gap with broadcast television.

At the last day of this year's IBC, the representatives of the companies that work on this project met up in Amsterdam, to come to a further agreement of the specifications that are part of the WAVE project.

Tools for validating CMAF content have been made available to members of MPEG.

One important issue that was discussed at length, were requirements for 'splice conditioned content', i.e. content formatted and signaled in way that guarantees frame accurate playback and allows for the straightforward insertion of other content, such as advertisements.

In addition, it was announced that the tools for validating CMAF content have been made available to members of MPEG. These tools are important, as they can be used to ensure that content is compatible with the CMAF specification.

Encoders and codecs

What would the world of media be without the constant promise of storing higher quality images in smaller files, using less time to do so? This year's IBC was no different. Intel hosted encoding technology from three competing companies on their stand. Along with Beamr there were:

  • Nokia Spain, presenting an Intel Quick Sync-based solution that focused on Adaptive Bitrate Streams (ABR), using content-aware encoding that varies the bitrate and resolution without quality loss.

  • MulticoreWare, demoing their x264- and x265-based UHDKit, which adds a proprietary layer on top of both open-source encoders to speed up ABR encoding by informing the process for one bitrate by the processes for the others.

Elsewhere V-Nova showcased PERSEUS 2, the newest iteration of their encoding technology, which is somewhat similar to Beamr's in claiming to offer additional compression using current codecs.

Of course, several big encoder vendors showed off their technical prowess as well, by demoing support for the much talked about AV1 codec, amongst other things. Depending on who you talk to, AV1 might prove to become a direct competitor or successor to HEVC in a few years.

Last but not least, there was a daring and surprising announcement by Swedish start-up Divideon. In an effort to monetize the dissatisfaction regarding HEVC's royalties, it presented a completely new codec, dubbed 'xvc'. Among its unique features are a publicly available source code and a licensing model that includes an indemnification clause to protect against third party infringement assertions.

Machine learning for everything

Without a doubt, machine learning is a buzz word that is often used to make current technology sound more smart and futuristic than that is actually is. At trade shows like IBC, this kind of techno-inflation is on full display as every company hopes to catch a bit of the spotlight while putting their newest products on display.

Techno-inflation is on full display as every company hopes to catch a bit of the spotlight.

Nevertheless, it's hard to deny that machine learning is trending in the world of video streaming. Companies like IBM and Amazon use it to power scene detection and to add metadata that describes the video image (e.g., which actors are part of a scene).

Or what about improving the process of ABR stream selection to increase the quality of experience for users when watching video? Right now, clients use fixed control rules based on simplified models of the deployment environment. Making these rules more flexible and incorporating performance observations of past decisions through machine learning, a more optimal performance across the broad set of network conditions might be achieved.

A final use case for machine learning that came up in conversations at the show this year was dynamic ad insertion, another buzz word that will immediately pique the interest of most people involved in video streaming. How can machine learning help? Well, it could smarten the decision-making process that decides what ad will be served to which person. Whether that will actually be useful? Maybe IBC 2018 will tell.