Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


HEVC/H.265 Update from Jan Ozer

Navigating upheaval in the codec wars

From about 2009-2013, the codec world was all about H.264. Sure, you heard about VP8, but relatively few websites other than YouTube used it, so its impact was minor. Starting in late 2013, you began hearing about HEVC/H.265, the standards-based codec for ultra-high definition (UHD) and beyond, with most pundits predicting a quiet and orderly succession from H.264 to H.265. In some markets, truth be told, that remains the most likely case.

However, the key factor that launched H.264 to the top of the streaming codec world—free, pervasive playback in Adobe Flash—isn’t happening for HEVC. In fact, while VP9 plays for free in more than 65 percent of notebooks and desktop computers, HEVC doesn’t play in any browser for free on any platform. Beyond that, a second HEVC patent pool launched, with exorbitant per-unit encoder/decoder royalties, and a proposed royalty on HEVC-encoded content, and there’s news of a third pool to come.

In September 2015, several open-source codec developers, including Mozilla (Daala), Google (VP9/10), and Cisco (Thor), joined with Microsoft, Intel, Netflix, and Amazon to form the Alliance for Open Media. This group will collaborate on a free, open-source codec to compete with HEVC, with the first “Alliance” codec expected to be released in early 2017. The Alliance codec is also expected to be the codec of choice for the Internet Video Codec (NETVC) group of the Internet Engineering Task Force (IETF).

Also during 2015, Google announced that YouTube viewers have watched more than 25 billion hours of VP9-encoded video. It’s still primarily YouTube using the codec, but mamma mia, that’s a lot of cat videos, and a lot of videos that aren’t encoded in HEVC.

The bottom line is that the relatively staid codec world is in a state of upheaval. Before you or your clients commit to any codec beyond H.264, you need to understand the players and the dynamics, which I’ll cover in this article.


Any discussion of HEVC has to begin with the cost side. As mentioned, there are two existing patent pools. The first was assembled by MPEG LA, which also formed the largest H.264 patent pool. The royalty terms for the MPEG LA group is $0.20/unit on encoders and decoders, after a 100,000 de minimus exception, with a U.S. $25 million annual cap, and no royalty on encoded content. The second pool is from HEVC Advance, which is backed by GE, and includes HEVC IP owners Technicolor, Dolby, Philips, and Mitsubishi Electric. HEVC Advance royalties range from $.80/unit to $1.50/unit for encoder/decoders, depending upon the playback device, with no deminimus exception, and no cap, plus a .5 percent content royalty on revenue attributable to HEVC. As you’ll see throughout this article, this is a huge change that dramatically changes the dynamic of HEVC usage.


In general, standards almost always win in markets like broadcast and OTT video; there’s simply too much investment to depend upon a proprietary technology, even if created by an alliance. Still, HEVC usage in these markets is expanding in dribs and drabs, so beyond the highly publicized Netflix, there’s simply not a lot of HEVC-encoded content.

The mobile market is also very standards-friendly, and HEVC playback is now included in Android 5 devices, and used by Apple for FaceTime on the iPhone 6, though strangely not available for general video playback. Of course, Android also supports VP9 playback, and YouTube certainly ensures that there is plenty of VP9 content available.

Again, with HEVC Advance, the cost of including HEVC in mobile devices has increased from $.25/unit with a cap of $25 million in the first year, to $.80/unit with no cap. With Apple closing in on 200 million iPhone 6 units (135 million in June 2015), this means Apple would owe a royalty of over $150 million to HEVC Advance, while the collective group of Android phone/tablet makers could owe the same or more. With the free Alliance codec coming, these costs could change the standards-friendly dynamic usually enjoyed by mobile devices.

Moving onto computer playback, though video consumption on mobile and OTT are burgeoning, there’s an awful lot of video is still watched on computers and notebooks, where HEVC playback is MIA. Though Microsoft originally announced that it would include HEVC playback in Windows 10, the company later pulled it, reportedly for concerns over performance. Again, when Microsoft originally decided to include HEVC, it was a $25 million dollar decision. Today, it could easily be a $1 billion dollar decision or more over the life of Windows 10. Given that Microsoft was a founding member of the Alliance for Open Media, you’d have to assume they are rethinking that decision.

In the meantime, while Apple has included HEVC playback on iPhones, it’s made no similar decision for Macs. In contrast, today, VP9 plays back for free in Chrome, Firefox, and Opera, which by some estimates, comprise as much as 70 percent of the browser market for desktops and notebooks. Microsoft has committed to adding VP9 playback to Microsoft Edge, the Windows 10 browser, so in the streaming market—which has been dominated by proprietary codecs in the past, like On2’s VP6—VP9 is much more strongly entrenched.

Significantly, while HEVC has been largely sitting still, over the last twelve months, Google has fully leveraged its investment in VP9, reportedly pumping out 25 billion hours of VP9-encoded video from April 2014 to April 2015. According to Google, VP9 can reduce video bandwidth by up to 35 percent, which my tests bear out. Given that the technology is free, you would expect any video distributor with significant bandwidth costs to consider using VP9 as well. Though no major publisher has announced their intention to date, perhaps the reality of content royalties on HEVC and the skyrocketing cost of encoder/decoder royalties might prove the decisive incentives.


In the meantime, alternatives to both codecs are coming. The biggest splash was made on April Fools day by London-based V-Nova, a four-year-old company, which claimed the ability to deliver “UHD in HD bit rates, HD in SD bit rates, and SD in sub-audio bit rates.” Most impressively, V-Nova has its first client, Sky Italia, an Italian Pay TV provider, that is using the codec for studio-to-studio contribution links. V-Nova is also reportedly working with a laundry list of other technology companies and video distributors, including Intel, Nvidia, Broadcom, and Hitachi.

In March, Austin, Texas-based COGO announced that it had trialed its new technology at the 2015 South by Southwest Music Conference and Festival, delivering a live concert streamed in 4K resolution. Precious little is known about the codec except that it is H.264-compliant, which presumably means that H.264 decoders can play the encoded video. The company claims to be able to produce “remarkable picture quality” at 3Mbps for 1080p and 6Mbps for 4K.

Most recently, in October, newcomer Tveon announced a new proprietary technology that “can deliver 4K-resolution UHD TV at below 2Mbps, and ‘true’ 1080p images at below 200Kbps.” Let’s be clear: Even if these quality-related claims are true, successfully marketing proprietary codecs is a very rough row to hoe. The largest companies are loath to invest in proprietary technologies, and many like Apple, Panasonic, Microsoft, Nokia, Qualcomm, and Ericsson have contributed technology to the HEVC patent pools, so they are invested its success. It’s always unclear when new technologies are stepping on existing intellectual property (IP) rights, so infringement claims are frequent.

While all this tumult is certainly interesting, technology gets adapted most quickly in boring times, where standards are in place and universally supported. Facing all this confusion, many companies, particularly content producers, are simply sitting still and keep pumping out old reliable, H.264.

Featured Articles