In the first part of this blog post series, we covered the history of TV viewing. Starting from the time when we only had one TV channel, we took a journey through the on-demand era towards the lean-back curated experience audiences expect today.
Today’s technology enables lean-back viewing experiences, such as creating virtual dynamic linear-type channels out of already transcoded and prepared video content. So, if you’re a content provider sitting on a great catalogue of content, virtual dynamic channels are a great way to further utilize your content and give your customers a great experience.
A virtual dynamic channel is a channel that is curated from several already compressed sources – VOD, live or a combination of the two. The purpose of the curation may be to provide a lean-back but relevant experience, or to boost viewing time by adding additional content before and after a live event. An important aspect of the concept is that this process happens in the distribution domain and not the production domain.
Instead of having a number of traditional fixed channels, it is possible to create virtual playlists, personalized or not, without the need for encoder capacity per stream or playlist. This introduces a valuable means to maximize relevance while keeping the number of decisions and interactions needed by the viewer down.
Use cases for virtual channels
In the traditional distribution model, the cost of creating a linear channel is high but the distribution cost is relatively low. With streaming, we have – more or less – the same cost for distributing linear channels as we do for VOD. By reusing the already transcoded content, virtual dynamic channels are a cost-efficient way to monetize content and attract viewers. You can generate hundreds of channels for a fraction of the cost compared to live 24/7 transcoding.
The cost for the service provider only exists when someone decides to watch something. This means that virtual dynamic channels, created on-the-fly, can be a very cost-efficient way to distribute channels with a smaller audience compared to using traditional distribution models.
This technique with virtual dynamic channels is a big success factor for the Free Ad-Supported Streaming Services (FAST), such as PlutoTV, Xumo, Tubi and Plex. They all aggregate a multitude of these virtual channels, insert targeted advertising and share the revenue with the content provider. These services are often preinstalled in TV sets or streaming devices and are now growing faster than subscription-based services.
There are different technical solutions to implement virtual dynamic channels, typically with various pros and cons. The solutions will differ in terms of their complexity to implement, dependency on other parts of the video pipeline and the experience they offer for the viewer.
1. Side-car playlist option
The first option – perhaps the easiest and most common way – is to create a side-car playlist (not to be mixed with the ABR manifest/playlist) which is provided to the client outside of the stream itself. The player will parse and play the assets one by one to achieve a curated experience adapted to the context.
There are obvious limitations, such as the client and player dependency, the restriction to VOD assets and its finite length. Once it reaches the end, it either stops or starts over. The experience is more of a looping playlist rather than a curated continuous flow. Since it only consists of VOD assets and does not have any actual time reference, a side-car playlist also offers the means to skip forward or backward, which might be good or bad depending on the use case.
2. Manifest Manipulation/live manifest option
Another common approach is to create a customized manifest or a new live manifest: a playlist in the adaptive bitrate domain, on-the fly based on a predefined schedule. The input can be adaptive bitrate segments that are stored on the VOD origin or segments created by a live origin, stitched together to create a live playlist and served through the content delivery network. This technique is similar to how server-side ad insertion is usually done where ad segments are stitched into an already existing manifest.
For this to perform well on the device or TV screen, it is important that the content is prepared properly and that the existing bitrates are as close as possible between the assets. As you would like to avoid transcoding, the closer in quality the segments are, the better. You still need to inform the player on the device when a change is done.
Managing gaps in the timeline
The technical implementation must also handle gaps in the timeline, usually by inserting slates, to avoid breaking the stream. If advertising should be inserted, the ad segments must be inserted in the same way as the rest of the content. Another way is to insert markers and handle dynamic ad insertion further downstream.
3. Pre-packaging Segment Stitching
The concept of manifest manipulation does not work for Microsoft Smooth Streaming (MSS) due to the principles of MSS. This could be managed if the segments are stitched together and presented as a linear channel prior to packaging. This method means that the stream can be packaged into any format, including MSS.
This method also means that segments can be renamed and a common encryption key can be applied to the entire stream, erasing all artifacts in the stream that reveals it has been stitched together from various sources. You will have fewer potential challenges with the players or ad-blockers using segment source or naming to decompose and compromise the stream.
On the other hand, if this is scaled up to very high number of personalized streams, it comes at a cost of cache storage in the CDNs since assets being used across multiple channels will be handled as unique segments. This could be mitigated if the CDN is aware of the structure of how segments are named.
Cord-cutting is accelerating across the world, making consumers turn to online alternatives. Instead of creating streaming alternatives that mimic traditional cable bundles, only served over the internet, there is technology that allows for more innovative services.
With virtual dynamic channels, you can generate hundreds of channels – or let’s call them playlists – for a fraction of the cost compared to traditional TV. What is stopping you?
By Magnus Svensson and Johan Bolin
Magnus Svensson is a Media Solution Specialist from Eyevinn Technology with long experience in small and large-scale technology solutions in the TV and Media area, as a leader and technical advisor in IPTV, OTT and streaming media delivery. Eyevinn Technology, an independent consultant company specializing in video streaming, distribution and technology. Eyevinn are helping large broadcasters, TV operators, newspapers and technology vendors to solve challenges with for example server side ad insertion, high volume live streaming, DRM protection and hybrid CDN solutions.
Magnus assignments includes streaming solution architecture, technical product management, business development and senior management advisory. Magnus is also involved in strategical assignments helping customers to navigate in the media industry, as well as conducting Technical Due Diligence projects.
Johan Bolin is chief product and technology officer for Edgeware, and has more than 15 years’ experience in the Telecom and Media industry. He has worked with both fixed, mobile and TV products both on the vendor (Ericsson, Edgeware) and operator (Tele2) side. With a passion for technology Johan has a wide and deep experience from Ericsson Research on subjects such as virtual and augmented reality and advanced video coding as well as managing the technology shifts from the product and content perspective on the operator side in both fixed, mobile and TV products. Today Johan leads Edgeware’s Product organization developing innovative and highly scalable products for distribution TV and video over IP and internet.
Fill out the form below and we will get in touch with you.