Encoding 2020: Experts Predict the Future of Video Encoding
Not surprisingly, the various vendors in the space see very different keys to success. For example, Ian Munford, Akamai’s director of product marketing, media services, sees the integrated storage, encoding, packaging, and delivery capabilities of Akamai and just a few other firms as key to meeting the needs of large content producers. “Where you store content is where you will deliver it from,” Munford says. “You wouldn’t want to move a 1TB 4K file from one service to another for delivery.” In Munford’s view, once a company decides to encode in the cloud, the vendor it chooses for storage and delivery will have a significant impact on who they choose for encoding, which will make it harder for smaller companies without the full suite of products and services to compete. He also noted that many customers simply want one number to call when problems arise and to eliminate inter-vendor finger pointing.
Jason Thibeault, Limelight’s senior director of marketing strategy (and Streaming Media columnist), agrees. According to Thibeault, Limelight’s Orchestrate platform (Figure 2) incorporates all the elements offered by Akamai, but also offers a content management system, and performs all encoding and transcoding on the fly, with no preconversion.
Thibeault says content owners want to publish as quickly and efficiently as possible and don’t want their content to bounce from service to service. He also believes that publishing is more important than encoding to his customers, so their choice of CDN will also effectively be their choice of encoding platform. “Our customers’ key performance indicators relate to how fast their video starts playing,” Thibeault says, “and how few rebuffers their viewers experience.” He sees an encoding environment dominated by Amazon, Limelight, Akamai, and Microsoft Azure, with many smaller encoding firms snapped up or otherwise disappearing.
As logical as Akamai’s and Limelight’s predictions sound, it flies in the face of the fact that DIY companies such as JW Player and Wowza appear to be thriving. During my discussions with Greg Twohig, JW Player’s product manager for media services, he shared that more than 2 million sites worldwide use the JW Player, delivering 9 billion video streams to 900 million monthly unique viewers, making JW Player the third most widely used player behind Facebook and YouTube.
Twohig asserts that their customers don’t want to be locked into using a subpar component of an integrated OVP service offering. “Successful streaming companies like Netflix, Hulu, YouTube, and Major League Baseball Advance Media are software companies,” Twohig says. “They prefer to create and manage the streaming experience themselves by assembling customizable technology components from different sources, rather than using the stock integrated services offered by a monolithic OVP.”
In the same vein, Brightcove started unbundling the components of its online video platform around 18 months ago. As David Sayed, VP of product management, explained, before the shift, Brightcove was a black box, where you uploaded your files at the start and grabbed an embed code at the end.
“We found that more and more customers just wanted a component of our technology, perhaps because they had already invested in their own content management system, or encoding capabilities,” Sayed says. “Now all major platform components are offered on a standalone basis accessible via mature public APIs.”
Albert Lai, Brightcove’s CTO, offered an interesting explanation for Brightcove’s strategy shift. “Companies transition through their own individual cycles, ” he says, “Some built their entire video delivery platform themselves, and realized that it wasn’t what they did best; their number one priority was to be a content company. So these customers now want a complete platform. Other companies are going through the opposite cycle, where they feel it’s essential to choose what they perceive to be the best solutions for all major platform components. Sometimes customers switch strategies from one approach to the other, and we needed a platform flexible enough to serve both strategies.”
Callahan also pointed out that signing on with a single CDN or OVP may limit leverage in pricing negotiations and may make it difficult to utilize other CDNs to serve viewers in areas where your selected CDN might not be that competent. He also agreed with many of the fundamentals of JW Player’s and Brightcove’s arguments, saying, “there are so many critical components to a video delivery stack—ingest, encoding, packaging and deliver, ad insertion, and player—that it’s tough to be best of breed at every one of those things.”
On the other side, Trow sees that many content producers, including companies such as the BBC, are frustrated with the level of effort necessary to deliver video on the rapidly expanding range of platforms. “Many large content producers,” he says, “want to simply hand off a mezzanine file to a service provider, and focus on their content.”
So the service provider market is headed in both directions, large-scale integration for companies who just want to hand over a mezzanine file, and discrete components for companies who want to customize their delivery stack to meet their own business requirements. What type of video will they be encoding and delivering?
Most vendors agreed with the general direction seen in the surveys. Callahan first mentioned that HDR is a significant issue for his company, which counts pay TV customers as a very large percentage of their business. Callahan related that there were multiple HDR technologies from multiple companies, many of which Elemental would have to support within their encoders. My discussions with Callahan were incorporated into an article called “HDR Is Here—But Don’t Rush Out to Buy a New TV Just Yet.”
In 2015, HDR encoding will almost exclusively impact those encoding for display on 4K TVs, and as Table 6 shows, by 2020, an increasing number of producers will target those devices. Delivering HDR video will require a complete upgrade of existing video production workflows, from cameras to editing stations to encoding. It almost certainly will require producing and delivering HDR-formatted video in multiple output formats. While you won’t have to support HDR to distribute video to smart TVs, if HDR is in your plans, it’s going to be a major investment and production hassle.
When will HDR hit computers and mobile? HDR-capable computer monitors are already available, primarily for video editors producing for HDR display. While prices are exorbitant today, that won’t always be the case, and cuttingedge streaming producers will start streaming HDR video to those devices. Someday, monitor manufacturers might agree on a standard before mass marketing their devices, or they might not. Certainly by 2020, there will be significant numbers of HDR monitors installed on desktops, and we’ll probably be into our second or third generation of mobile HDR-capable displays.
The other up-and-coming technology is virtual reality, which delivers an immersive experience to viewers watching via a virtual reality headset. Imagine watching Game 7 of the NBA Finals from a courtside seat (in your own living room), so you can move your head to follow the action, gaze up at the scoreboard and replay screen, and even scan around to see how your fellow Warriors fans are rejoicing. That’s the promise of VR, and once implemented, it will completely revolutionize how we watch live events.
At a high level, VR works by taking feeds from multiple cameras and stitching it into an experience that looks cohesive on a headset. Multiple feeds mean massive data input and very significant video compression. Though there has been much investment in this space—as an example, in 2014, Facebook acquired VR vendor Oculus for $2 billion—standards and applications are far from being set. By 2020, however, it would be surprising if multiple VR applications have not been commercialized.
Scalable Video Coding
It’s fitting that we conclude our look toward 2020 with a blast from the past, scalable video coding, or SVC, which I called “the future” in an article posted back in early 2009. Briefly, SVC delivers an adaptive streaming experience using a single file, divided into layers that improve quality, frame rate, or resolution. Though SVC never got the headlines, according to Romain Bouqueau, CEO of GPAC Licensing, a provider of open source software, SVC was deployed in multiple high-profile applications, including Skype before the Microsoft acquisition. SVC has also been built into multiple products offered by video collaboration vendor Vidyo, who contributed scalable coding extensions to HEVC.
Bouqueau sees hybrid delivery of content as the big use case for SVC. For example, an HD resolution file could be one price, with higher resolutions and frame rates available at an additional cost. SVC would allow CDNs to store a single file for all delivery configurations, which would save disk space and improve caching. So while SVC isn’t yet a checklist feature on any encoding program, it just might be by 2020.
For me, the biggest surprise from the survey was how many respondents have already switched over from Flash to HTML, plus how quickly 4K and larger videos will be in the mix. Also of note is the relatively high percentage of VP9 video planned by most respondents, which is getting much more traction than VP8 ever got.
On the vendor side, it looks like a fascinating battle between many Davids and a few Goliaths, which should ultimately come down to how many clients consider themselves in the software business versus the content creation business. HDR promises to be every bit of the mess it is today, and VR a tantalizing look at how we’ll watch many live events 10 and 20 years from today.
This article appears in the September 2015 issue of Streaming Media magazine as “Encoding 2020.”
Encoding.com's annual report shows that H.264's solid lead is growing, while HLS is dominant in adaptive streaming. Flash Video will be gone in two years.
With Reserved Cloud, broadcasters and others trade metered usage pricing for a fixed monthly fee and their own reserved instance.
Learn the best practices for reacting and responding in real-time to issues in an encoding and delivery workflow to ensure a seamless experience for viewers.
While Bitcodin is still in development and has a few rough edges, it displayed impressive performance and quality in our testing.
Encoding for VOD is very different from live transcoding. Here's what you need to know to find a service that serves your needs.
When an encoding job is highly demanding, it takes powerful gear to create and judge all the files. Here are the right tools for the job.
Companies and Suppliers Mentioned