The State of Video Codecs 2015
It’s thus unlikely that a substantial portion of the installed base will want software-only HEVC playback because of the power drain and associated reduction in battery life that come from playing UHD-encoded video without hardware acceleration. Compared to the desktop and notebook markets, where large swaths of the installed base can become available overnight, growth in mobile UHD playback will largely be prospective only.
While custom services targeting these newer devices will likely sprout in 2015, the short-term mobile UHD market doesn’t seem to inspire complete library conversions of general-purpose web content.
Before discussing OTT/Smart TVs, let’s identify several realities. First, YouTube is transcoding its entire library, particularly 4K videos, into VP9 and has announced no plans to support HEVC. If you want to watch 4K YouTube videos, it looks to be VP9 or nothing. Second, cord-cutting is a widespread phenomenon, meaning that traditional broadcast viewing via cable and satellite is on the decline. Third, according to a knowledgeable source, adding VP9 decode to a system on a chip (SOC) or other chip already configured with HEVC hardware decode could cost as little as $0.15. This for a device you could sell for as much as $40.
Given these realities, would you create a chip, STB, OTT device, or smart TV that didn’t offer VP9 decode, even if only for YouTube playback? Probably not, but that doesn’t make the inclusion of VP9 in a rash of products introduced at the 2015 CES from companies such as Broadcom, Sigma Designs, MediaTek, and Philips, symptomatic of a larger trend toward the use of VP9. As a streaming producer, it also doesn’t mean you should choose VP9 over HEVC.
Judging from the format support offered by leading on-prem and cloud-encoding vendors, there doesn’t appear to be an outpouring of demand for VP9 encoding among premium producers that would supply content for UHD distribution deals. I scanned spec sheets from a number of providers: Elemental, Envivio, Harmonic, Digital Rapids, Telestream, and Thomson Video Networks, all of which sell in bunches to broadcast companies, movie studios, and similar companies. All currently offer HEVC encoding. None of them offer VP9 encoding.
I spoke to marketing types at these and other companies. Off the record I heard: “We don’t have many requests for VP9 support at this time and no firm plans to add it in the near future,” and similar sentiments.
On the record, Elemental Technology’s CMO Keith Wymbs says, “Elemental is constantly assessing next-generation codecs. Today, we see significant demand for HEVC while we see limited requests for VP9. We’ve enabled more than 100 companies so far with HEVC capability as they prove out their ecosystems and some have moved to commercial deployment. Because of our software-defined approach, we can implement VP9 when it makes business sense and we see a broader market need to do so. The CE announcements at CES suggest that VP9 could pick up steam, but only time will tell.”
Of the major cloud vendors, Encoding.com supports HEVC but not VP9, and won’t unless it sees demand for it, which it hasn’t to date. Zencoder currently supports HEVC but not VP9, while Amazon supports neither.
Clearly, the aforementioned announcements at CES do not presage a surge in VP9 demand. In this regard, it’s instructive to understand that Google’s interest in VP9 is very different than that of the typical streaming producer. Specifically, given the volume of video encoded, stored, and delivered by YouTube, bandwidth costs are obviously a significant concern, and the ability to develop, customize, and continually advance a high-performance codec to meet its precise workflow needs can produce very significant cost savings. Having to work within the constraints of a standards-based codecs would be antithetical to those needs. In addition, with YouTube, Google has the single largest source of video on the internet, a big enough asset to convince CE manufacturers to support a new codec, especially if it costs less than a quarter.
While bandwidth is an issue for all large streaming producers, codecs are more a necessary evil than strategic tool. Most producers want an economical, easy-to-use codec that plays everywhere so they can expand their libraries, find new viewers, and expand their business. So far, at least in the mobile and OTT/smart TV markets, the leading candidate is HEVC.
Obviously, this segment depends totally on hardware, which means tiny, tiny quantities throughout 2015. There will be content deals in this market in 2015, but unless you have one, converting to either UHD format doesn’t seem warranted.
Computers and Notebooks
Where HEVC has the edge in mobile and OTT/smart TVs, the situation is reversed with traditional computers and notebooks, where VP9 playback is available today on Chrome, Firefox, and Opera, which represents a 62 percent browser share, according to W4Counter. No existing browsers support HEVC playback, though Microsoft has announced that it will include HEVC in Windows 10, which is scheduled to ship in mid to late 2015. Since Apple has included HEVC in the iPhone 6, and Google with Lollipop, you would assume they would support HEVC playback in their respective browsers, but neither has made any announcements.
Going beyond the browser, HEVC playback is available in much smaller numbers in the DivX and VLC Players, which can be used as browser plug-ins. Adobe has announced HEVC support in Adobe Primetime, its Flash-based premium distribution platform, but has not announced that it will support HEVC in the free version of Flash. So, for today, the number of VP9-capable computers currently dwarfs that of HEVC-capable computers.
For Google, and perhaps other large publishers, a browser-specific solution is sufficient for a complete library transcoding. To explain, in a WebRTC Update meeting held on June 30, 2014, and published on YouTube, Google reported that it was transcoding its entire library to VP9. It also stated that VP9 was the default format for Chrome and that 60 percent or more of all daily playbacks on Chrome were using VP9 (Figure 2). Obviously, the storage and bandwidth savings resulting from VP9 more than justified the costs of the encode.
On the other hand, for other producers, targeting a single browser is not enough. To target all the browsers, you would need to create or license a custom player that could use either codec. According to discussions with Ittiam and Fluendo, developers of multimedia applications for computers, mobile devices, and other platforms, the user experience would vary based upon the selected codec and technology, and might require that the user download and install a plug-in. Both companies are in this business; Ittiam offers a player and HEVC and VP9 codecs, while Fluendo typically designs its solutions around a framework called GStreamer, which includes HEVC playback.
Here, costs could be substantial, particularly if you’ll need to convert from Flash to HTML5 and/or the Media Source Extensions (EME). While most of the pieces of this conversion are in place, some aren’t, particularly if the conversion requires DRM. For example, no one is quite sure whether Apple will license FairPlay for use on Safari via the Encrypted Media Extensions, though Apple may have done so to Netflix. So not only will you have to work through the new encoding workflows, you may have to develop around standards that are still very wet behind the ears. You’ll also have to build or license a player and decoder, then pay the HEVC royalty to MPEG LA.
Though MPEG LA’s stated policy is no royalty on HEVC encoded video distributed over the internet, until and unless Apple, Google, and Microsoft supply HEVC playback in their respective browsers, distributors who choose to distribute HEVC-encoded video will have to pay to include HEVC decode in their player. Those who choose to supply HEVC decode to Firefox and Opera will likely have to do so forever, since neither is likely to ever license HEVC.
Is this market worth investing in? If your strategy depends upon shipping 4K content to computers and notebooks in 2015, then obviously the answer is yes. If you’re considering UHD for the bandwidth savings, use web analytics to see how much of your bandwidth is spent on delivering to computers and notebooks. If it’s 60 percent, add 60 percent to your bandwidth expense—you should save about 30 percent of that total by using either UHD codec.
Figure out what that number will look like for each of the next 3 years, then add those numbers up. If the total is worth chasing, then call your IT department and start assessing costs. Factor in that you’ll be converting from Flash to MSE over the next couple of years anyway, and that Apple, Google, and Microsoft will likely add HEVC decode to their browsers in the next 12–18 months. Be sure to consider the pros and cons of Adobe PrimeTime, which will extend the life of your Flash Player into the UHD era.
Will it be as seamless as switching from VP6 to H.264? With PrimeTime, it should be; with all other technologies, almost certainly not. But you’ll have to make the Flash-to-MSE/EME and HD-to-UHD transitions sooner or later, and the clock is ticking on the first year’s bandwidth savings.
[This article appears in the 2015 Streaming Media Industry Sourcebook as The State of Video Codecs.]
Licensing fees may stall HEVC adoption, while AV1 is nowhere to be seen and VP9 is on the rise. Will 2017 be at eventful as the previous two years?
After a year of uncertainty, HEVC seems poised to finally make inroads, if not for 4K then for bandwidth reduction.
Which delivers better quality, encoding time, and CPU performance—HEVC or VP9? We put them to the test to decide once and for all.
Which codec delivers better image quality? Which is more compatible? And what about Daala, the spoiler codec currently being developed from scratch?
Unlike the quality of either WMV or VP6-based Flash files, which are relatively homogenous due to their common origin, the quality of H.264 video varies depending upon who developed the codec and, to a lesser degree, the encoding tool. Here's a comparison of H.264 files produced using codecs from Apple, Dicas, and MainConcept.
Tues., Aug. 26, by Jan Ozer
Allows for Downloading and Streaming at Various Rates
Companies and Suppliers Mentioned