Streaming Learning Center Where Streaming Professionals Learn to Excel
Related Articles
When I first heard that Google Chrome played HEVC, I checked the weather in Hades to determine if it had, in fact, frozen over. Reports were inconclusive. Then I started reaching out to colleagues and contacts whose opinions I value to get their reactions and thoughts. Many spoke on the record, and many didn’t. In this article, I’ll attempt to piece their inputs together to help you understand exactly what happened and allow you to draw your own conclusion as to what it might mean.
In case you haven’t heard, here’s what we learned from the Bitmovin blog entitled Google Quietly Added HEVC Support in Chrome. Daniel Weinberger, who penned the post, stated, “Quietly, without any announcement or updates on support pages, Google fixed a bug in Chrome with a significant implication for the video streaming industry: Support for adaptive streaming of HEVC/H.265 video content has finally been enabled!…It’s now officially supported for Chrome 104, and with a little investigation also found out that it’s enabled by default for Chrome 105 for all platforms, ready to be used in the wild.”
Note one major caveat, however, that HEVC is only enabled in Chrome when playback support is already available on the underlying platform, whether Windows, Linux, Mac, iOS, or Android. However, most of these platforms, particularly Mac, iOS, Android, and Windows, already support HEVC in hardware. There are several other caveats, particularly relating to DRM and HDR, which we’ll explore below.
Just for fun, if you want to see if your version of Chrome version plays HEVC, navigate Chrome to bit.ly/play_hevc on the Bitmovin site and see if the video plays. Then come back and finish the rest of the article.
Contents
CanIuse or Not?
By way of background, HEVC, the designated standards-based successor to H.264, launched in 2013 to great acclaim and even greater expectations. But three patent pools and over-reaching and poorly defined royalty expectations created significant antipathy, which is largely credited with the 2015 launch of the Alliance for Open Media (AOMedia) of which Chrome developer Google was a founding member. The AV1 codec shipped in mid-2018.
One of AOMedia’s key strengths is that members own all major browser platforms and can decide which codecs to support, and which to ignore. Before the Chrome announcement, other than Apple, which started supporting HEVC before it joined AOMedia, no other members supported HEVC in their current browser.
When I checked CanIUse (https://caniuse.com/?search=hevc) for the percentage of HEVC support in May 2022 for a Streaming Media presentation, the number stood at 18.53%. This compared to 74.6% for AV1, which was two years younger, but benefited from support from all AOMembers in their current browsers except for Apple. A few days after the Bitmovin article, HEVC support skyrocketed to over 80% before it dropped back to 22.91% (on October 14, 2022), prob ably because HEVC playback won’t be enabled by default on all platforms until versions 107-109.
Beyond the browser, HEVC support is nearly universal. Buoyed by 4K and High Dynamice Range (HDR) content, HEVC is the codec of choice for premium content in the living room. Apple added HEVC support to iOS, macOS, and tvOS in 2017 while Android 5 and later versions also support HEVC, mostly via hardware support to preserve battery life and ensure full-frame playback. This enabled HEVC support for Android apps but not Chrome, so publishers that distributed via browsers couldn’t easily deploy via HEVC.
With a global market share that approaches 65.52% on computers and mobile devices, Google’s refusal to support HEVC in Chrome made HEVC implausible for producers that distributed primarily to browsers. As a result, HEVC dominated on Smart TVs and dongles and had some success in mobile, but that was it, despite offering significant bandwidth savings over H.264.
Now that Chrome plays HEVC this bottleneck is gone. This raises several questions, like will this open the floodgates for HEVC encoding, and will it slow the momentum for AV1? These questions were among the many that our respondents answered.
As a caveat, we weren’t able to get any input fom Google for this article. I asked a high-level Google contact to introduce me to a Google employee who could speak on the record, but after agreeing to do so, he never responded. If Google does get in touch after publication, we’ll add some comments to this article online.
Let’s start with a look at what we know about Chrome’s HEVC implementation.
About Chrome’s HEVC Support
As most readers know, when you connect to a website to play a video, the website queries the player to identify its capabilities. For this reason, most (if not all) producers that distribute advanced codecs like HEVC, AV1, or VP9 also encode in H.264. If the player can’t decode the advanced stream, the website sends a manifest for H.264. For this reason, you don’t need 100% coverage to start distributing HEVC to browsers. Chrome will be able to play a significant percentage of HEVC videos, but there are some limitations.
Most playback capabilities are delineated on the Chromium Github site here. By way of background, Chromium is the open-source project upon which Chrome (and Microsoft Edge and Opera) are based. Back in 2015, Bitmovin reported a “bug” in Chromium entitled “Enable HEVC support in HTML5 MSE on HEVC-enabled Android phones.” On September 19, the status of the “bug” was quietly updated to “fixed” with no explanation.
What we learn at the Chromium Github site is that:
Leverages existing playback capabilities– The Github document is entitled “A guide that teach you enable hardware HEVC decoding for Chrome / Edge, or build a custom version of Chromium / Electron that supports hardware & software HEVC decoding.” As you can see in Figure 1, the page offers three downloads. Two are for Chrome and Edge (Mac) and both of these “Support HW decoding only.” There’s also a version of Chromium (e.g. Not Chrome) that appears to support both hardware and software decoding. Whether a software decoder is included in the download is unclear, but it seems clear that Chrome itself needs a hardware decoder to play HEVC.
In answer to the question “Will HEVC decoding be enabled in Chrome by default in the future?” the Github document states “Chrome >= 107.0.5300.0 has already enabled HEVC HW decoding support for ChromeOS, Mac, Windows and Android by default, Chrome >= 108.0.5354.0 also enabled HEVC HW decoding support for Linux by default. Chrome 107 release version will be available after 2022-10-25.”
In terms of significant limitations, the Bitmovin article states that “The biggest drawback is that HEVC with Widevine DRM is not supported at this point, only clear, unprotected content. It’s unclear whether Google has plans to add support for this in the future or not.” Several sources have confirmed this which I’ll elaborate on below.
Lack of DRM support is a big deal since premium content protected by DRM comprises a huge chunk of existing HEVC-encoded video. As one anonymous source commented “Google not enabling HEVC+Widevine in Chrome is like having the cake but not being able to eat it.”
Regarding HDR, the Github document shares the following table. Regarding Dolby Vision, the document states, “PQ backward compatibility single layer Dolby Vision (Profile 8.1, 8.2, 8.4, although when using API query dvh1.08.07, it still returns “not supported”), not support IPTPQc2 single layer Dolby Vision (Profile 5), not support multi-layer Dolby vision, not support Dolby Atmos audio (E-AC3).”
Fortunately, incomplete Dolby Vision support isn’t as significant as the lack of DRM because most producers create Dolby Vision and HLG formats anyway to support the broadest range of HDR-capable devices.
Taking a step back, it’s interesting that though Chromium is separate from Chrome, the Chromium blog often speaks for Chrome with statements like “ “Chrome >= 107.0.5300.0 has already enabled HEVC HW decoding support for ChromeOS, Mac, Windows and Android by default.” I asked Bitmovin’s Wenberger about this and he explained, “Chromium is the open source project Google is maintaining…So if there is software support implemented, every browser (like Chrome) can still decide if they want to ship this feature or not (if a flag for it exists).”
Given that the bulk of what we actually know comes from the Chromium blog, and not directly from Google, there’s still some risk that Google might say “never mind,” and disable HEVC support. I don’t think that’s likely but it’s definitely worth remembering.
Will We See a Huge Update in HEVC Usage?
So, those are the basic facts; what does it all mean? Will this spark a run on HEVC encoders and encoding services? Responses ranged from a strongly positive endorsement to a muted meh. For example, Alex Giladi, a Fellow at Comcast, stated “We are excited about the support of HEVC in Google Chrome, which is an important step forward in ensuring ubiquitous access to a more efficient video codec. This will help to drive better, more efficient video delivery, as well as better streaming experiences for viewers around the world.”
Several other large producers stated that they would send HEVC to any platform that supported it (DRM permitting), and several service providers/encoding vendors like Harmonic and Bitmovin were recommending the same to their customers. Speaking for Bitmovin, Daniel Weinberger commented, “If you are already doing HEVC in your workflow, add the option to your offering on all platforms your application supports! Any good player will choose more efficient codecs if supported on a given platform or falls back to a less efficient one that’s supported (like H.264 typically).” Harmonic’s Thierry Fautier, Vice President, Video Strategy stated, “For our customers delivering UHD on PCs, they now have full access to HEVC and HDR, which is a huge benefit.”
Noted compressionist Fabio Sonnati, Media Architect, and Encoding and Streaming Specialist at NTT Data, added “for sure H265 on Chrome could help to reuse the H265 renditions that many OTT streaming already produce for devices like Firestick (4K), connected TVs, etc… an enlargement of H265 usage and audience may even subtract use cases from AV1.”
Answering the question “Will this accelerate usage of HEVC for producers who aren’t using it yet?” another streaming media consultant responded, “I think so. People are ready, waiting, EXCITED for the next generation of video codecs to become a reality and everyday element to a degree. This is a step closer that may push those waiting on the line to jump in or at least take another baby step.”
Still, several other respondents didn’t think that Chrome supporting HEVC would bring a rush of new HEVC-encoded content. For example, Greg Ellis, COO & VP of Business Development and Sales at Dacast, commented that “The near-term impact is muted somewhat because of the current viewing device demographics.” As you can read in his complete comments below (all respondent’s full comments are available at the end of this article), Greg was noting that most premium content is consumed on Smart TVs and mobile devices which already support HEVC playback, the latter via apps. According to Ellis, “That leaves business users and video advertising IMO as the areas that will be most impacted.”
Magnus Svensson, a Media Solution Specialist at Eyevinn Technology, added “I think that HEVC support in Chrome will make it easier for services to move to HEVC as the codec is now supported in the majority of the browsers. But I’m not sure that we’ll see a huge increase just because of this. Lack of support in Chrome has been a factor in holding back, but I don’t think that this has been the deciding factor.”
One limiting factor is the lack of DRM support already cited, plus the fact that even once DRM is available, that browser-based viewers will only see relatively low-resolution versions. As one video engineer who works for a large OTT shop commented, “The sad part is… HEVC is great, but most people get SD or 720p in Chrome due to DRM limitations. So it’s not like it’s going to make the biggest difference (given that HEVC really starts to shine at 1080 and up given all the big transforms etc).”
Will There be Content Royalties?
Beyond DRM, the continuing potential for content royalties is concerning to several respondents. Behnam Kakavand, the video R&D lead engineer at Evolution Gaming, commented that HEVC playback in Chrome “does increase the likelihood of using HEVC for us. However, it is only one factor, and there are others at play as well, like the additional costs including royalty, which require investigation from our legal department. In fact, the content royalties do have a big influence on the decision on whether we are ever going to use HEVC or not.”
To this, Harmonic’s Fautier added, “While Chrome’s support of HEVC will help it to be deployed on PCs, the royalty issues have not been resolved.”
I asked patent attorney Robert J. L. Moore from Volpe Koenig about the potential for content roylties. He pointed out that “neither of the two existing pools, which account for about 85% of existing HEVC patents, is charging royalties for streaming. The Velos members, by and large, have either joined the Access Advance HEVC pool (Panasonic and Sony) or have their own robust licensing programs that major implementers, including content providers, are probably licensing already.”
Continuing, Moore stated that “certainly, there exist HEVC patent owners unaffiliated with pools, and content providers not already licensed to the unaffiliated owners’ portfolios could be required to pay some royalties. Still, in light of the above, and in light of the fact that we haven’t seen a significant amount of HEVC SEP enforcement so far against content providers, I don’t see Velos’s exit significantly changing the royalty paradigm for content providers implementing HEVC. In short, I expect that content providers will most likely continue to enjoy the ability to implement HEVC for low or no royalties.” Every content company considering distributing HEVC encoded content should obviously perform their own due diligence, but Moore’s comments seem to indicate that the potential for royalties is slight.
One comment from an anonymous contributor who works for a streaming technology provider wraps this part of the discussion in a nice bow. He stated “Although Chrome browser is not the biggest in terms of viewing share (compared to TVs, STBs, native Android/iPhone, …), the HEVC support in Chrome is another step in simplifying the question of which codecs/profiles to deploy, to reach as many clients as possible. This could push operators towards AVC (as ‘default/fallback’), and HEVC for higher-resolution/premium content, even more. Given that at least one large broadcaster spontaneously referred to the change in a conversation about profile/codec selection, and that it seemed to impact their decision, I’d say it is a meaningful change…”
What About AV1?
Another interesting topic is the impact of Chrome’s playing HEVC on AV1. Is codec support a zero sum game where if HEVC gains AV1 loses? Our respondents were mixed.
Harmonic’s Fautier commented, “AV1 has its own ecosystem, driven by GAFAN companies. This announcement will not derail their plans to avoid HEVC.” Bitmovin’s Weinberger added, “the position hasn’t really changed. AV1 support is also already quite widespread – more than 70% according to caniuse, it’s more efficient than HEVC and it’s royalty-free. Jumping on the same royalty-free claim, Evolution’s Kakavand added, “Given that the royalties on content are still unclear for us, we still put our bet on the AV1 as a viable next-gen CODEC. If it turns out that the content royalties are not a blocker, then we can think more seriously about the possibility of using HEVC for desktops.”
Of course, AV1’s royalty-free status is far from certain, as Sisvel’s AV1 patent pool attests. And, at least one prominent analyst, Alex Davis, Senior Analyst at Rethink, thinks that HEVC’s gain might be AV1’s loss. To explain, Davis feels that Google refused to support HEVC to promote AV1, and opines that Google supporting HEVC now means ”that the list of priorities has changed inside Google. The big juicy spiciest take would be that this is Google admitting defeat and opening the door for HEVC, but it could easily be a new leadership within the Chrome project – either somewhere at the top, or lower down on the bug program…Part of the argument for AV1 was that HEVC was not workable in its use cases, so now that is seemingly no longer the case.”
The Big Picture
Beyond the tactical implications discussed above, some of the most thoughtful observations came from Eli Lubitch, president of codec company BEAMR, during a rambling hour or so long phone call. I mention the phone call because most of this will be paraphrasing from hastily typed notes, with few direct quotes.
To start, Lubitch thought that Google’s move to support Chrome related to the fact that the most ubiquitous class of capture devices-phones and tablets-almost all had the option to capture in HEVC format (many as the default). Since the primary browser for viewing such videos is Chrome, HEVC support was an obvious step if long overdue.
Beyond the decision itself, Lubitch lamented the low profile, quiet release by Google relating to this critical decision. Google, a prominent AOM member, is now widely embracing HEVC. So, nine years after the launch of HEVC and five years after the AV1 shipped, “the evidence cannot be reduced to a clear strategy. The uncertainties this creates slow the adoption of both codecs because “customers are reluctant to move when the ecosystem isn’t well defined.”
He also noted that while large companies like Google have the resources and talent to support multiple technologies like AV1 and HEVC, many smaller companies don’t. “Smaller companies can’t duplicate or replicate progress along two major front lines, they have to choose. And if they choose wrong, it may be very costly.”
Finally, Lubitch observed that while the video market used to quickly choose a winner between competing technologies – remember Betamax vs. VHS, or Blu-ray vs. HD-DVD – the codec market accelerates in the opposite direction, not only open source vs. standards-based but VVC vs. EVC vs. LCEVC. Clearly, this is suboptimal for all participants, from codec developers to device manufacturers, to content publishers, and especially the viewer.
So, this is where we are. I had hoped to deliver clarity but there was simply none to be found. As Lubitch so aptly stated, “the evidence cannot be reduced to a clear strategy,” and Google itself isn’t talking.
