Encoding 2020: Experts Predict the Future of Video Encoding
What’s the encoding world going to look like 5 years hence? That’s what we wanted to know, and we decided to investigate from two different angles. The first was via a survey taken by 335 Streaming Media readers, the second from conversations with vendors in the streaming media space. Let’s start with a look at the survey data.
For the record, we solicited responses via an article on the Streaming Media website, offering a $100 Amazon gift card as an incentive. After some demographic questions, the survey focused on technology and implementation questions, gathering data about current practices and planned practices for 2020. You’ll see the data in the seven tables that follow.
One of the demographic questions asked about annual spending on encoding-related products and services. I thought it would be interesting to see how opinions varied between those at the top of the spending ladder ($500,000 or more), and those who spent less than $50,000. So for each reported survey response, I include data for all respondents and for respondents in these two categories.
The change section on the extreme right of each table reports the changes from 2015 to 2020 in all three categories, and it’s been color coded to make it easier to spot significant trends. Any negative change of 50 percent or less is colored in yellow; beyond that, it’s colored in red. All positive changes up to and including 100 percent are colored in blue, while those exceeding 100 percent are colored in green.
Table 1 presents where respondents are encoding VOD files now and where they plan to encode in 2020. As you can see, the big encoding spenders are leading the charge to the cloud, already encoding 41 percent of their files in the cloud. By 2020, all three categories are close to 60/40 in favor of the cloud. The green numbers in the change column reflect this momentum. If you thought on-premises encoding was going away, however, it looks like you’re mistaken, with 39 percent of all encoding anticipated to be on-premises in 2020.
Table 2 shows the same numbers for live encoding. Interestingly, though live cloud transcoding has only been available for 2 or 3 years, its numbers are very close to VOD encoding. Big spenders again lead the charge over other respondents, though the numbers are very close by 2020.
Table 3 shows codec usage data for 2015 and 2020. Glancing at the color coding, it’s clear that VC-1 and H.264 are the big losers, while H.265 and VP9 are the big winners. Though only 10 percent of respondents planned on using VP9, that’s still a huge number compared to current VP9 usage, and it’s telling that the big spenders anticipate 10 percent of their streams to be VP9. HEVC stealing share from H.264 is obviously a surprise to no one.
How quickly will the world adapt 4K/8K? Just check Table 4. By 2020, close to 40 percent of all video distributed by our respondents will be 4K or larger. Not surprisingly, big spenders are again the first movers with 4K and 8K video, though by 2020, even the hoi polloi will have caught up at 4K, with 29 percent of their streams anticipated to be at this resolution.
On to Table 5. By 2020, most pundits would predict that DASH would be the dominant adaptive streaming technology, which Table
5 confirms, though there’s other intriguing data. For example, the big spenders dumped RTMP-based Flash much faster than other respondents, with higher percentages of Smooth Streaming and HDS than the respondents as whole. However, the red boxes in the change column shows that big spenders are dumping these technologies in favor of DASH much more quickly than the other groups.
Also interesting is the predicted continued usage of HLS, which probably reflects the general consensus that that Apple might never adopt DASH. In addition, Flash might be dead in some users’ eyes, but it looks to be the technology of choice for 20 percent of streams in 2020.
Where will we be sending these streams in 2020? Table 6 shows the stream allocation between computers, mobile, and OTT/ smart TVs now and in 5 years. Computers are dropping, but not off a cliff, and OTT/smart TVs, currently a stronghold for the big spending crowd, will become an increased focus for all streaming producers, particularly the smaller ones, with an increase of 141 percent.
Finally, let’s take a high-level view of the Flash to HTML5 transition, which Table 7 shows is a transition that many respondents have made with almost half of their streams. Perhaps some day, Adobe’s enduring plug-in will be an afterthought, but apparently not by 2020, when 22 percent of the largest respondents expect the streams distributed to computers and notebooks to be watched via Flash.
That’s what our readers thought, now let’s check the vendor side.
The Vendor Side
When we announced the survey on the Streaming Media website, I invited all vendors who wanted to chat to get in touch. Some did, and I reached out to some others in critical markets and market segments, including individuals in on-premises and cloud encoding companies, CDNs, and OVPs.
The first area of inquiry was the shift of encoding to the cloud, which we already know is a significant trend. Five years ago, cloud encoding and on-premises encoding were two separate markets; today, most on-premises encoding manufacturers have a cloud offering. To a degree, Elemental Technologies led the charge, signaling the shift with a white paper issued in early 2014 titled, “Software-Defined-Video: A Game-Changing Framework for the Video Marketplace.” In the paper, Elemental predicted the demise of ASIC-based encoding in favor of software that could run anywhere, including on traditional encoding appliances, in the cloud, and in private clouds. Elemental led the charge to the cloud and is one of the leading cloud encoding vendors.
Other companies have made similar moves. When I spoke with Ian Trow, Harmonic, Inc.’s senior director emerging technology and strategy, he said that Harmonic had previously discontinued ASIC-based encoder development in favor of developing for cloud and virtualized environments. In early 2014, Harmonic announced VOS, a virtualized media processing architecture designed to run on public and private clouds (Figure 1). Beyond encoding, Harmonic has combined multiple other functions that used to require custom hardware, such as graphics and playout, into a software architecture that can run on commercial off-the-shelf (COTS) hardware. These software-based architectures have multiple advantages, including scalability with ebbs and flow in demand, and the ability to share hardware with other IT functions.
Trow sees multiple trends pushing large media companies to the cloud, whether private or public. Some broadcasters use traditional hardware for their linear broadcasting, with cloud encoding for internet-delivered videos. For these companies, as younger viewers consume increasing quantities of video online, it increases the importance of cloud encoding to these companies. Other media companies simply prefer the OPEX over the CAPEX model, a traditional argument for cloud encoding. “Five years ago,” Trow says, “most of our customers were 100 percent on-premises encoders, today they’re probably 90–95 percent eon-premises. By 2020, I would expect them to be 60/40 in favor of on-premises.”
Beyond these considerations, Greggory Heil, chief encoding officer at Encoding.com, sees many other reasons why companies are moving to the cloud. For example, according to Heil, the first generation of cloud encoders essentially implemented “FFMEG in the cloud,” and lacked parity with on-premises encoders, both from a features perspective and APIs. Now that gap has closed, particularly with Encoding.com, which has focused on adding key workflows such as Dolby Digital Plus and the Nielson Watermarking Encoding Engine. Encoding.com can also run Harmonic’s VOS on its cloud platform, and encode using presets from Harmonic’s Carbon Coder. (Harmonic owns shares in Encoding.com.)
Beyond the maturity of cloud encoding platforms, Heil says large media companies are now storing their media content in the cloud, overcoming their initial security-related concerns that prevented many from encoding in the cloud. In addition, faster bandwidths and acceleration technologies have dramatically reduced upload times, while vendors entering the cloud space have reduced the costs of storage and compute. Interestingly, Heil says that Encoding.com will also license its software for use in a private cloud, which will help the company add features such as live linear playout, which doesn’t particularly perform well in a public cloud environment.
What functions aren’t moving to the cloud? Though Telestream is rapidly expanding its cloud offerings, CTO Shawn Carnahan noted that there is significant transcoding and other processing that occurs prior to and immediately after production, including “converting camera formats to edit formats, and edit outputs to mezzanine for distribution, much of which might not be a good candidate for cloud deployment.” So companies performing production as well as distribution will likely have to retain significant on-premises encoding/transcoding capabilities to support these pre- and postproduction capabilities.
The Integrated Cloud
Beyond evolving into full-featured encoding products, the cloud encoding market has become more integrated over the past 5 years, and will continue to do so over the next five. In 2010, for example, there were cloud encoding companies, off-the-shelf player vendors, OVPs, and CDNs, and all almost exclusively performed a single function. Now cloud encoding companies have distribution arms, such as Bitcodin’s Bitmovin OVP (powered by the Bitdash player). Off-the-shelf player vendors, such as JW Player, have an OVP (JW Platform), while CDNs such as Akamai and Limelight provide OVP-like services, including player support with mobile SDKs.
In September 2014, Elemental Technologies, once a standalone appliance-based encoding vendor, launched Elemental Delta, a video delivery platform that functions as a just-in-time packager and origin server. According to Elemental’s senior director of product marketing Mike Callahan, once Delta runs in the cloud, which should occur soon, Elemental will be “a player away from a full service OVP.”
Carnahan sees “more and more encoding being done as part of a service that can handle encoding, packaging, and distribution. We see ourselves getting into the delivery business at some point.”
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