Hacker News new | comments | show | ask | jobs | submit login
The Fight for Patent-Unencumbered Media Codecs Is Nearly Won (ocallahan.org)
495 points by bzbarsky 11 months ago | hide | past | web | favorite | 170 comments

HEVC is basically mandatory in new TV sets in Europe as part of the new DVB-T2 https://en.m.wikipedia.org/wiki/DVB-T2

The new digital TV standard frees some frequencies for 5G. Most channels will broadcast with HEVC Main 10 (bits). There is also a Main 8 which will probably not be used for long and people should be careful about what they buy.

At least we won't have HEVC on the Internet.

The new North American ATSC-3 TV standard also uses HEVC so that particular royalty will be charged on a TV purchase in that part of the world. Dunno how much it will end up adding. The current practice seems to be to pack in as much chargeable intellectual property into these "industry standards" as is humanly possible.

wait, is there really forward progression in broadcast OTA television in the US?

I've always thought that cable television and modern broadcast television was all still mpeg-2. HEVC would be a huge leap forward in tech (and backward in openness, I suppose?)

Side-step in openness. All MPEG-2 has in its favor its the patents expire earlier.

iOS 11 also encodes videos into HEVC, and photos into HEIC, which is basically HEVC still frames. Considering the network effect of iOS devices, this fight is hardly won.

The VP9 decoder installed base is about double HEVC's:


So to me that bodes well for future AV1 support in all kinds of devices.

It doesn't matter. No one is going to support a video format that's not compatible with iOS.

Adobe effectively lost the war with Flash on mobile the day that Apple decided not to support it. Even though it was available for Android.

Both YouTube and Netflix support VP9:


And YouTube doesn't do 4K encoding in H.264 anymore. If you want 4K video from YouTube you need VP9 support (or AV1 when YouTube starts AV1 encodes):


The subject of the article was the "fight for patent-unencumbered media codecs is almost won". To "win" that fight means that content providers can abandon patent protected formats. Netflix is still having to deal with patents and now they had to dedicate more resources to support an additional format.

Large content providers like YouTube or Netflix really don't care about the patent and have the resources for storage and encoding multiple formats. Where patents hurt are smaller entrants that would benefit from patent free formats.

They still have to support patented formats and now they have to spend money on encoding and storage on yet another codec if the want to use VP9. What's the benefit of supporting two formats? One that is supported everywhere and one that isn't it?

> To "win" that fight means that content providers can abandon patent protected formats.

And that's what's happening. VP9 is royalty-free and AV1 will be royalty-free.

> Where patents hurt are smaller entrants that would benefit from patent free formats.

So it's good that both YouTube and Netflix support VP9 and AV1. It helps make royalty-free video formats commonplace on the web.

> What's the benefit of supporting two formats?

It's worth it for the bandwidth saving. VP9 significantly outperforms H.264.

And that's what's happening. VP9 is royalty-free and AV1 will be royalty-free

Who are these content providers that are "abandoning" H.264?

so it's good that both Netflix and YouTube support VP9 and Av1

Good for who? End users don't care if content is encoded in a patent free format. They just care about the content being available on their platform.

> End users don't care if content is encoded in a patent free format.

This is a tired argument. The same argument was attempted when VP8 was new, and when VP9 was new, and now again with AV1. I'm sure the same, worn-out ground will be revisited when AV2 comes along.

End users don't care that TCP\IP is royalty-free or that HTTP and HTML are royalty-free or that any of the other commonly used formats and protocols are royalty-free. They don't care because they don't understand the issues.

Developers care. Builders care. I care. I want to implement video on the internet for any use case without having to consult a lawyer just to understand the licensing implications of doing so. VP9 and AV1 make that possible. VP9 and AV1 normalize video on the internet by making it royalty-free like all the other internet formats and protocols.

Content companies and hardware companies and software companies don't join AOMedia for the fun of it. They join because it's the practical choice:


There may be a few hundred thousand of developers that do but there are hundreds of millions of users that just want their stuff to work.

Profit is in the users, not in the developers

But you've got to produce before you can consume and you've got to deliver to the consumer. AV1's licensing reduces costs and AV1's lower bitrate delivers better picture quality earlier with less bandwidth.

The choice isn't AV1 or users. It's AV1 and users.

Products already exists. That's what users consume. That's also why users don't care about technological developments unless those technological developments improve user's experience, not make it more difficult for the users.

How does it "reduce costs" if you still have to support both?

And I can't find how much licensing actual costs for H.264.

If there is a demand, content will come, I doubt the difference in cost made by the licensing fees are going to mske a difference.

No, but potential for bandwidth caps with fees for going in light of the removal of network neutrality legislation might tip the balance in the favor of the codec that looks better at a lower bitrate.

What does that have to do with patents and why people should care?

No, but potential for bandwidth caps with fees for going in light of the removal of network neutrality legislation might tip the balance in the favor of the codec that looks better at a lower bitrate.

I think the problem is that it limits competition. The established companies with deep pockets have an advantage if royalties have to be paid.

This is bad for users, whether they realize it or not. They end up with fewer (and potentially worse) options.

The cost for a creating, hosting and streaming video is huge. If a new entrant has the funding to create and host compelling video, I doubt that the minuscule cost of royalties will make or break them.

The positive affect on large content providers is non existent when they have to support both and negligible as far as cost even when they do.

It doesn't really have an affect on small content producers. I do live streams for a non profit. We pay one time for Wirecast and the Black Magic hardware. Pay a neglible cost to send the livestream to our provider and they broadcast it. iOS users at least can view from the browser using HLS. We also send another stream to Facebook. How does H.264 patents affect us?

You don't have to "consult a lawyer", unless you are creating your own codec. You pay the licensing fee which in the grand scheme of things is lower than all of your other costs - hosting, development tools, internet costs, etc. and you're done.

Note that even if H.264 may be free for your use case[1], HEVC is not. For HEVC, you must pay per title or subscriber, if you are content producer.

[1] Which I wouldn't be so sure. Many H.264-using devices come with disclaimer, that they are not licensed for professional use and that you may need your own license. As usual, consult your lawyer. The point of VP9/AV1 is, that this is not necessary.

> End users don't care if content is encoded in a patent free format. They just care about the content being available on their platform.

End users do care about the cost of their devices. Patent-free formats lower that cost. And patent encumbrances just make it more difficult and costly for manufacturers to include support; sometimes some formats get left out due to cost.

I certainly agree that end users don't care about patents, but they care about some of the practical problems that patent-encumbered formats cause down the line. These are usually not showstoppers, but they help cause customer confusion and sticker shock.

Having to support both formats in the short- or medium-term is definitely annoying and costly for manufacturers, but if we do actually win the open format war (I think the OP is a little overly optimistic there), that problem goes away.

A lot of UGC is video. If uploading (or just storing) your videos in an industry-standard and very efficient format for free can get you in hot water, or require shady downloads of cracked / banned software, users are going to be unhappy, and would cry and protest. Remember the early days of mp3, or DivX / XviD, banning of VLC, etc.

You get a much wider and more loyal user base if you're smart enough not to enforce such silliness anymore.

H.264 is a legacy codec from the VP8 generation and is only needed to support legacy platforms. Current Internet encodes for modern devices should be VP9 (nobody's going to step into the HEVC patent pool minefield), and future encodes should be AV1.

End users pay hundreds of dollars in royalties to companies for their electronics. Get rid of royalties and the technology race to the bottom will pass those savings on to consumers.

"Hundreds of dollars"? Source?

I don't think that's out of line.

I was working on networked audio/video devices back in ~2005 when that space was pretty new. We had a line-item BOM cost for encoder/decoder licenses that was not chump-change, which got passed on to customers.

I can only imagine costs have gotten much worse since then. When you take into account all the electronics your average connected person owns: phone, tablet, TV, AppleTV/GoogleTV/Roku, computer OS, etc., it really starts to add up.

Even if the parent poster is overstating and it's on the order of $50 per person, that's a ton of wasted money when you consider the total cost.

MPEG-2 used to cost $2.50 per device. MPEG-4 is another 25 cents or so. Add another $2.00 for HVEC. MP3 was $0.75-$5.00 per unit. Each video you stream from Netflix or Hulu costs a few cents royalty. Royalties for such things used to be even higher before the threat of unpatented codecs dropped prices (eg, h.264 becoming free to compete with vp9).

Consider all the electronics you've owned over the past 20-25 years. You undoubtedly paid a dollar or two for less complex ones and much more for more complex devices.

>The subject of the article was the "fight for patent-unencumbered media codecs is almost won". To "win" that fight means that content providers can abandon patent protected formats.

Eventually they will, the only missing piece for AV1 to become the de facto successor to h264 was Apple, with them onboard AV1 will cover all of web and mobile. I mean look at the companies behind AV1:

Google, Microsoft, Apple, Amazon, Netflix, Facebook, Mozilla, Cisco, Intel, AMD, ARM, IBM, NVidia, Broadcom, and that's about HALF the list.

Google chrome plays 4k YouTube videos on the Mac but Safari does not. Is there a reason that Google cannot ship the VP9 codec with their Apple TV app? Or is the Apple TV 4K hardware unable to accommodate VP9 and will never support it?

>> No one is going to support a video format that's not compatible with iOS.

You might try reading the article next time. Apple recently joined the Alliance for Open Media. That indicates to me that they are likely to support AV1. Apple was really the last holdout and it would be poor business for them not to play.

How does Apple joining the alliance negate the sentence that market share doesn't matter if a format is not compatible with iOS?

My contention wasn't that Apple wouldn't support VP9, it was that if Apple didn't support VP9, it's not likely that content developers would leave iOS users out in the cold.

Phones are only used for a few years. Within 4-5 years, virtually no one will use a HEVC-only iPhone.

What about smart TVs?

They make a fine display device to plug an updated Apple TV, Fire TV, Chromecast, or Roku into.

Software can be updated if the hardware support is there. Arguing that new codecs don't have hardware support in older TV's has nothing to do with how patented that codec is. Speculatively adding support for upcoming codecs is easier if you don't have to pay to play.

What about them? Even if we have to wait 10 years or so to phase out support for those that don't support newer formats and can't be upgraded, that's still a win.

By then it won't matter. The patents on h.264 will have expired.

So what? Then we have to deal with the newer h.265 patents, and the even newer patents associated with whatever comes next.

Pushing open formats now means we develop a habit of only allowing open formats into industrial standards, which means in the medium and long term we all win with lower costs.

iOS users are welcome to purchase devices that have proper support for standards at their next refresh cycle.

How did that workout for Adobe with Flash?

Most users did not want flash. Most users don't want web apps or apps inside web views, but they'll use those if they have to. Why switch phones to get flash if you really wanted a native experience in the first place?

"Most users didn't want Flash". That's a lot of revisionist history. Apple was being knocked for not supporting Flash from the time the first iPhone was released in 2007 until a year or two after the iPad was released in 2010.

Rephrase, most users didn’t want phones that burned up their battery capacity, or more often just out and out crashed repeatedly when playing videos.

Indeed. We witnessed ethis with H264, where even the likes of Nokia (before they sold the phone business to Microsoft and later licensed off the brand) leaned in on not having webm be part of the HTML video standard.

This because they had a bunch of devices out there that could do H264 in hardware, but had to do webm in software, and thus playing the latter format would take a greater toll on the battery.

And even if every damn device use a programmable DSP these days, the problem is not solved. Because only the OEM can be expected to provide the firmware needed for new formats.

Effectively the only way to win this would be for GPLv3 or similar to be made law, so that consumer devices are not beholden to OEMs for maintenance.

Yes, it matters. Apple joined AOM, which means it will kill HEVC in the same way it killed Flash. Once the HEVC-only devices are out of the market, Apple will say it will drop support for HEVC.

I think you really have to look at installed or planned hardware decoders, or closely coupled software/GPU decoders to really declare victory.

Edit: software/GPU decoders, not hardware/GPU decoders.

At least for HEIC, it's not really exposed to the Web - pictures are silently transcoded to JPEG when uploading. Apple could swap out HEIC for AV1 behind the scenes whenever they want, or alternately sliently transcode HEIC to AV1.

Did you read the article? Apple is going to support the new standard. It doesn't matter that the old phones support HEVC.

Now if only Apple would join Vulkan, too, so we can all agree on a cross-platform graphics API, too. Because of Apple Khronos now has to make a subset of Vulkan that will be cross-platform. This means that if developers end-up adopting that one over Metal on iOS, the situation will be worse off for everyone, including Apple.

Apple has gotten better with supporting standards from the Steve Jobs days, but I think it still has some way to go.

Apple has joined AOM is different to Apple announcing support.

They joined AOM as a top level (governing) member, why on earth would they do that unless they are going to support the codec on their software/hardware ?

Outside of the office walls in Cupertino, who knows?

What does it cost to join as a governing member? Money? Apple has an unbelievably big pile of it. Spending it relieves their tax burden, so why not buy the most expensive level of membership?

Cynically, maybe they want to influence it so that it sucks, so while Apple does support it, Apple's going to have their own proprietary code, so Apple videos on Apple gear looks great. Everyone else gets the AOM code which, for some reason, everyone else's looks (and sounds) ever so slightly worse.

They could influence the codec so it works better on Apple hardware, so they get an advantage while designing the GPU that will go inside the iPhone XI or XII.

They could use this to see what the codec will support, and then decide not to support it out of spite, just like the iPhone X does't have USB-C, even though their laptop team has discovered that technology.

Point is, why knows? Yes, it is reasonable to think that Apple joining AOM means they'll put AOM codecs on their devices, but unless you work at Apple at the executive level, it's pure conjecture (and even then, executives are allowed to change their minds).

AV1 will likely be voted as IEEE 's NetVC. Much like opus.

Apple joining simply means they won't sue AOM's company for video codec patent.

But why would they join if they're not going to support (as in use) the codec ?

They joined AOM at the highest level, which means they get a seat on the board of directors, I can't see any reason why they would do this unless they have decided to use it on their products.

I am only guessing, from what I know of Apple. So dont take my word as concrete information.

Apple uses Opus too, I think IETF announced it as a RFC standard. ( If my memory serve me correct )

But Apple only support the usage of Opus in limited ways.

So why would Apple not use it? If we forget about the patents and royalty fees for the moment. AV1 isn't necessarily better then HEVC from Apple's prospective.

HEVC is actually a standard, and has been for many years. There are many difference implementation of HEVC encoder and decoder, all adhering to this standard and conformance. Even Apple has their own implementation, there are Open Sources like x265, and many other commercial solution as well. AV1 isn't a standard, ( yet ) and there is one and only one implementation. Now before any geeks starting arguing about software benefits of implementation as standard over written standard, you need to think of this not as a pieces of software ( While it certainly is ), but a tools for professionals. Those who do TVs, and movies, etc all wanted something different and they want choices.

Google, or On2 to be precise; continuing its tradition of over hyping its codec. AV1 at its start was no where near 30% better then HEVC. And since they claim AV1 is now another 30% better then their initial version, you would have expect it to be 1.3 x 1.3 better then HEVC. But truth to be told AV1 when tested by panel of expert provide tangible better results then HEVC, at the cost of much slower encoding time. While that was done a while ago and AV1 has since improved a lot more, we shall wait and see when it finalized.

Everyone has been saying they will be forced to support AV1 because of Netflix and Youtube. Well that is only half correct because Netflix are already encoding in HEVC. And Netflix, has yet to provide their newer results of HEVC vs AV1. My guess is that they are waiting for AV1 to finalize. VMAF is a lot better then all other PSNR or SSIM, but it is still not perfect.

There are zero Full hardware AV1 decoder ( There may be partial Hardware decode that could be enabled with update, but you be the judge how many hardware manufacture is going to enable that then to sell you something new ) , but there are billions of devices capable of decoding HEVC already.

So apart from being royalty free, AV1 doesn't have a lot of advantage to it. The question is, for most business, will using AV1 saves me more money in the long term then using HEVC.

Apple joining on board, with No Press Release from AOM, and only Cnet manage to pick up, suggest this may be an Intentional leak. Or one way of Apple saying, HEVC Groups, you either do as I say and lower your fees, allow free software decode implementation ( One of the group already allows that ), or you know what? You can hold on to your patents for as long as you want and you wont earn a single dent in the billion of iOS devices in the future.

>HEVC is actually a standard, and has been for many years.

Once the AV1 bitstream is frozen, it will also be a 'standard', only this will be royalty free.

>there are Open Sources like x265, and many other commercial solution as well.

Nothing prevents x265 devs from making xAV1 (or whatever AV1 will end up being called), their spokesperson over at Doom9 has already said that they will go where the market goes in terms of encoder development, they are also very pissed in regards to the HEVC licensing debacle.

>we shall wait and see when it finalized.

Indeed, the bitstream is (supposed) to be finalized this month, after that happens we will finally see optimization take place (basically rewriting all hot spots into handwritten assembly) and thus be able to assess the quality claims and just how much slower it will be.

>Well that is only half correct because Netflix are already encoding in HEVC.

Netflix was one of the first companies to join AOM to develop a royalty free codec, it seems clear their intentions is to replace HEVC with AV1 once wide hardware support arrives.

>There are zero Full hardware AV1 decoder

There can't be until the bitstream has been frozen, from what I've read the first hardware supporting AV1 will be 12-18 months after said bitstream freeze.

Throughout the development of AV1, there's been constant consulting with hardware developers and they have had a large say in how AV1 works, the hardware companies that are part of AOM are:

Intel, AMD, ARM, Broadcom, NVidia, Realtek and now recently Apple

>So apart from being royalty free, AV1 doesn't have a lot of advantage to it.

If the estimates are correct, ~30% better compression is a HUGE advantage, another advantage from the point of the companies in AOM is that developing a codec themselves means it will fulfill their needs much better.

Google was going this route ever since they purchased On2, but now it really has reached critical mass with their third generation codec (VP10) being the base for AV1 which has made practically all the big tech companies come together and solve their codec needs using it.

>Apple joining on board, with No Press Release from AOM

I think Apple didn't want to ruffle any feathers with MPEGLA and the other HEVC licensees, Apple picking up HEVC support across their products was seen as a good sign for HEVC, the same Apple joining all the other tech giants in AOM backing AV1 is quite the opposite. Thus they join with no fanfare. Pure speculation of course.

By then time hardware support for these codecs come out most of these iOS devcices will be obsolete.

Apple won't switch codecs again shortly after going the HEVC/HEIF route. All new iPhones will stay there.

I think Apple's hand will be forced by major content providers. Hulu, YouTube, Amazon, Facebook, and Netflix all want AV1. YouTube and Netflix in particular want to adopt AV1 quickly. Netflix has been talking about supporting AV1 by the second quarter of this year:


They won't remove support for HEVC/HEIF for quite some time, but they will support AV1 alongside them, why else would they join AOM at the top level (governing member) ?

This means AV1 will be the de facto next generation codec standard on the web as it will be supported by all browsers and all mobile devices, something HEVC never will.

HEIF itself is an image container format that supports a number of different image formats (like HEVC, AVC, and JPEG):


Apple could and probably will add AV1 support to their HEIF implementation in the future.

Ultra HD (4k) Blu-rays use HEVC as well, so we will definitely not get rid of that standard any time soon.

Are UHD disks going to be anymore than the modern version of laserdisk?

It seems to me that even bluray is basically dead at this point. The younger generation isn't buying huge stacks of movies to fill out the 200 disk shelves they got for christmas, and the older generation frequently doesn't even have a bluray players.

Random (couple year old) link about bluray vs dvd sales numbers.


Optical disks are becoming more relevant for me as Netflix's catalog shrinks because content owners want to launch their own streaming. If I want to watch something today and in 6 months, I better buy it on disc, because who knows if it'll still be available later. Also, the minivan can play blu-rays, but streaming netflix on I-5 in central california doesn't seem likely to work.

Well, i'm in the same boat with respect to netfix and disks. But my usecase tends to be more along the lines of watching the movies on my tablet in the gym (which requires time shifting them if you will).

That isn't possible with UHD blurays at the moment, and I've sworn off buying DRM encumbered products. Just one of my reasons is that exactly 100% of the HDDVD's I purchased that were warner brothers products have bitrotted (its apparently true of nearly all of them if you believe other peoples posting) despite the fact that the disks from other studios continue to play. Some people have gotten the disks replaced with blurays, but WB didn't even respond to requests I made.

So, I won't buy DRM encumbered video products. I might rent them, but they can kiss off if the price is more than redbox or similar.

Really, what I want is to just buy an unencumbered digital version that isn't screwed up.

Give me something like the steam experience but without even the "DRM" of that.

Also set a reasonable price point. 5 USD for a SEASON of a great show sounds like a good high price starting point. Make it inexpensive for me to GIFT seasons to others as well. Maybe 2.50 for the first season to someone else, and then lower in cost after that.

How to increase profits? Stop the lobbying, consumer-fighting-disservice, and DRM insanity and just focus on making a good product and delivering it in an EASY way for consumers to get.

$5 USD is probably too low, depending on the length of the season; $2 USD / disc probably covers manufacturing and distribution (for Blu-Ray; DVD manufacturing maybe $1 USD), but you need something to pay for the talent too.

Clarifying, download + standard license to the copyrighted content in all formats in perpetuity only.

The specs on DVB-T2 are pretty crazy, at least in Germany. 1080p 50fps 10 bit, that is a lot of hardware that is obsolete.

HEVC is H265, Netflix already uses it with 4k clients that can support it.

True, but Netflix is one of the founding members of AV1, which is what they want to replace HEVC with.

Broadcast TV is rapidly dying, though. The very fact that they're changing codecs around to "free some frequencies" for more valuable spectrum users is the proof of that.

No, that's just the general progression of frequency allocation as we develop better coding and modulation schemes to more effectively use frequency.

On the analog side all new voice specifications on VHF/UHF are mandated to use 12.5kHz instead of the old 25kHz wide channels for the same reasons. Amateur radio is currently the only one I know of that still can use 25kHz wide FM frequencies[1].

P25, DMR, DStar and Fusion all use 12.5kHz since they can get digital audio into much narrower spectrum(DMR in fact supports two channels over one 12.5kHz channel using TMDA which is pretty awesome).

[1] http://wiki.radioreference.com/index.php/Narrowbanding

Also, as second point, broadcast TV is still very relevant for rural areas where high speed internet access hasn't made significant inroads. It's pretty straightforward to get a VHU/UHF antenna up on a roof/tower and get a reasonable number of channels.

It's anecdotal but I also know of a fair number of cord cutters who supplement their Netflix/Amazon/Hulu with local broadcasting.

I only see the HEVC mention for Germany and France. I'm not really seeing how it's "mandatory for Europe."

Anyways, for the countries that did make it mandatory, this is why it's stupid to mandate a protocol or specific technology into law. Haven't they learned from South Korea's IE/banking mess?

There is no law mandating that, it's just what you need to support if you want to decode the signal you actually receive with your antenna. The broadcasters have agreed to a common standard (which makes sense, imagine the mess if everybody broadcasted with a different technology).

Legacy TV is dying out anyway, so eventually, some manufacturers would start making TVs without HEVC that would be targeted for Internet video and it will only snowball.

I don't believe this is accurate. Do you have a source?

> Do you have a source

Sure. So called "cord cutting" trend has been around for a while already, and legacy TV has been losing users in millions.

See: https://duckduckgo.com/?q=Study%3A+Cord+Cutting+site%3Ahttps...

TL;DR: legacy TV is dead already, it just doesn't know it yet (or rather they already know it, but want to squeeze as much money as they can, before they hit the dust).

Note that DVB-T2 is a terrestrial broadcast standard (albeit not used in the US) and terrestrial broadcast is free/low cost in most places, hence US-focused articles about people dropping cable because of high monthly costs has little effect on the success of DVB-T2.

When it does come to pay TV, note that pay TV is still growing in Western Europe, mostly in places where they've historically had very little marketshare (v. terrestrial broadcasting).

Sure, broadcasting is still part of the legacy TV, but as the rest, its decline depends on the growth of good quality Internet availability. In the long term it will be gone, and radio will be used for Internet itself. Clearing of spectrum for that purpose is already gradually happening and other radio based Internet access types are coming as well (low orbit satellites and etc.).

Cord cutting typically refers to PAID TV, not OTA. OTA has seen increases because people are getting away from cable.


Tunerless TVs are definitely happening, but they are all smart TVs that need HEVC to support Netflix/Amazon/etc.

Netflix and Amazon are founder members of the AOM that is developing and promoting AV1. If the smart dongle or TV supports AV1 they'll preferentially use that to get better quality at lower bandwidths.

Netflix and Amazon are both moving to AV1, so while that's true today, it won't be for long.

Smart TVs need to support VP9 for YouTube 4k. Netflix works with VP9 too, so HEVC doesn't seem necessary.

I don't think the patent issues are at all resolved or easy. Variousparties hold patents and have no intention of designing a product around them and simply exist to extract rent from others.

They will wait till the format gains traction and then sue the players with the biggest bank accounts.

Most companies understand this, and will accordingly take it slow, preferring to deal with the known risk of existing codecs.

> I don't think the patent issues are at all resolved or easy.

Same applies to any codec, including HEVC. I.e. patent trolling can always happen and HEVC users aren't immune to it either. So AVC1 wins being actually officially free.

HEVC has been around longer, and has been widely deployed, so the risk is much smaller.

"HEVC has been around longer, and has been widely deployed, so the risk is much smaller."

Isn't this the opposite of what you said in the parent?

"They will wait till the format gains traction and then sue the players with the biggest bank accounts."

Let me try again: HEVC has been deployed by big players with deep pockets. Therefore, any patent trolls that might have targeted it have already been lured out of the woodwork.

A new codec like this however, has an unknown risk attached, because no one knows how many trolls might be lurking out ther e.

You could have said exactly the same thing about VP9. Tons of FUD over its patent status, but nothing stopped its adoption.

You could say exactly the same thing about any new codec before it's widely adopted.

As I said in my post, AV1 is more strongly protected than previous codecs because anyone who sues over AV1 can't use it themselves, so only pure patent trolls can sue over AV1.

Anyway this is only a short-term concern. Probably as soon as this year Google, Netflix and others will start deploying AV1 and we'll see what happens.

> HEVC has been deployed by big players with deep pockets. Therefore, any patent trolls that might have targeted it have already been lured out of the woodwork.

Depth of the potential targets' pockets was never a deterrent for patent trolls. On the contrary, they prefer to go after the big fish and they have nothing to lose (except their patents). So back to square one. HEVC is as risky as anything else plus it bears already declared huge tax.

Patent trolls can decide to attack any time they feel like it.

Free codecs actually are in somewhat better position, because they do due diligence to work around known related patents. And attack from unknown / invalid can always happen.

So overall HEVC is nowhere less risky, and on the contrary, many explicitly declare upfront, they want money for it which is its major downside.

This situation seemed almost hopeless 10 years ago. Very glad to see things have changed so much for the better.

I wonder how much this is enabled just due to patents expiring. Patents (theoretically anyway) last 20 years, so you'd expect 10 years to see off a lot of them. I don't know so much about video, but in the end the solution to mp3 patent encumbrance was mostly to just wait for the patents to expire.

All the important MEPG-2 patents are gone, which is a big help.

The reason for the traction of unencumbered media codecs lately is that MPEG LA got greedy with the license terms for HVEC, and now the cost of licensing HVEC/H265 is more than a lot of big licensor's are interested in paying. This was not true during the H264 period, and the reason there wasn't nearly as much push for open codecs then. So in a roundabout way we can thank the shortsightedness of MPEG LA, for the coming era of open codecs.

Reminds me of the old OPEC strategy of not making oil too expensive, lest they push industrial nations to develop alternatives.

It is not just MPEG LA BTW.

It's strange that the Alliance for Open Media hasn't put out a press release about Apple joining or added Apple's logo to the "about us" page:



AOMedia seemed to do that quite quickly when Facebook joined. Maybe Apple asked them not to, but I don't see why Apple would.

I found this odd too, so I was skeptical of the claim, but I found this sentence on the home page[1] that mentions Apple as a founding member:

> Founding members are Amazon, Apple, ARM, Cisco, Facebook, Google, IBM, Intel, Microsoft, Mozilla, Netflix and NVIDIA.

A similar sentence exists on AOM's about page[2].

[1]: http://aomedia.org/ [2]: http://aomedia.org/about-us/

It's also been widely reported in tech news and Apple has not denied it.

This might be a bit off-topic but what is Cisco's interest here?

Video conferencing.

The Thor codec tech they contributed is all about focusing on that use case, real-time encoding etc.

They also previously contributed the h.264 plugin to Mozilla (they already pay the capped royalty amount so it's effectively free for them) to ensure their hardware components would be able to interoperate with popular browsers using WebRTC.

Any word on the current state of encoding speeds? From what I remember of VP9, it was very, very slow in comparison to e.g. H.264/H.265.

VP9 is getting a lot better. Since the advent of the -frame-parallel and -tile-columns options, encoding scales well with CPU cores. The only frustrating thing for me is these aren't used by default. The VP9 encoding guide has info on their use: http://wiki.webmproject.org/ffmpeg/vp9-encoding-guide

On my i7-2600k, encoding a Blu-ray takes about 10-12 hours for H.265 or 18-20 for VP9 with handbrake's default 1080p settings for each.

So yea, quite a bit slower, at least on my old hardware.

I still go with VP9, though. I don't buy that many movies, and leaving it running in the background doesn't seem to make a big impact on gaming.

I should also mention that my BIOS has a performance setting that can be "Energy Saving", "Normal", or "High Performance", and I keep mine on "Energy Saving".

Kicking it up a notch or two could undoubtedly boost the performance a bit, and a manual overclock should go even further. (My chip is stable at 4.7Ghz, but it consumes way too much power to leave it there all the time - the computer idles at ~70 watts on "Energy Saver" mode!)

I jumped from your chip to a 6 core Broadwell-E and it seems to give me about 15% more performance out of x264 encoding and 5% from x265.

I see the same jump from C4 to C5 Amazon machines. I'd love to chat about encoding with you at some point.

Yikes! Is that 5-15% normalized to per-core performance, or does it really get almost 0 benefit from having 50% more cores?

That difference from C4 to C5 makes more sense given that they each have the same number of cores available and it was a smaller generational gap.

Send me an email sometime if you'd like - nathan@[my username].com

As a reply to other threads that ask if this means Apple will switch to AV1: encoding speed is the number one reason presented as a benefit of HEVC in one of the WWDC talks on HTTP streaming. So... I'd guess they'll still record things in HEVC, but maybe add support for decoding AV1.

Encoding speed right now is: either you have a computer cluster or you'll die before the encoding is done. But it will probably get better.

It's true that the encoder is very slow at the moment, but it's also true that Bitmovin managed to produce a live AV1 stream 9 months ago:


They also have an AV1 demo page which works with Firefox Nightly:


If they haven't tried optimizing encoding speed yet, how can they be sure there isn't some misfeature in the standard that will hinder high-speed encoding? Like IIRC with some old On2 codec, where the order of the tokens in the encoded stream was not ideal for an optimized implementation.

They don’t care. It’s Google, Netflix, and Amazon and their concern is recurring bandwidth, not onetime cpu costs.


"We will be satisfied with 20% efficiency improvement over HEVC when measured across a diverse set of content and would consider a 3-5x increase in computational complexity reasonable."

Some of the Alliance members want to use it for video chat (e.g. Cisco, Google) or live streaming (e.g. Twitch/Amazon, Google) so it'll get done at some point.

I can see their point (and that of any other content provider) in that context, but I'm sure they do care. It would be fine even if it required a dedicated hardware encoding bloc, but lack of reasonable RT encoding (including low wattage) would create a significant technical impediment to full general adoption, so I can't believe that optimization hasn't been at least somewhat considered. There is of course still plenty of real time streaming that's important, though in those cases reduction in quality would often be an acceptable tradeoff. However, Apple and Google both have a quite reasonable and absolute need for real time (or a close enough cheat), high quality offline encoding on mobile systems within the physical constraints of a handheld form factor. End users will also want to (again, quite reasonably) be able to immediately share what they make without the need for a lengthy reencoding stage or trip through a middleman.

So if there wasn't at least a timeline for 4k+ RT on iPhone or Android devices I can't see how HEVC wouldn't stick around for a very long time. That a live stream has already been demonstrated should mean that this has been considered and they do care though, and thus won't be a problem. Ideally of course at least the high end of current generation devices would be capable of AV1 recording through software updates, which would enable moving on much more rapidly from HEVC. For that at least though I think it'd be an acceptable tradeoff to have those become legacy if that's what it takes to get us into a patent-unencumbered future afterwards. A pity software patents couldn't have just been eliminated entirely but I guess that remains politically infeasible for now.

It's not onetime CPU costs, at least for Google. Netflix and amazon don't process that much video (relatively), so they can afford slow encode times, but youtube gets tons of video uploaded every minute, they want fast and cheap encoding.

For each individual video, it is a onetime CPU cost, but not a onetime bandwidth cost.

For ongoing operation of the platform, at scale, you might argue that it's an ongoing CPU cost as more and more videos are added constantly, but the bandwidth to distribute those videos goes up at an equal rate on average.

I don't know any actual numbers, but it seems very plausible to me that a vast increase in CPU cost for a minor decrease of bandwidth cost could still be a welcome trade-off.

If you're making an argument that availability of CPU power at that scale is a concern, it seems like YouTube has the option of Google's cloud services as an excellent candidate for massive, interruptible compute power.

It's not just bandwidth cost. Google suggests their big win with VP9 deployment is reaching people across the globe with relatively powerful computers but terrible internet connections. By using more powerful compression (paying for it with more computing work at both ends) they can get higher quality video to these users.

And from a blunt business perspective: Higher quality videos = more watch time = platform dominance = more ads viewed.

> vast increase in CPU cost for a minor decrease of bandwidth cost could still be a welcome trade-off

I don't think that this would be true, though. Recall the 80-20 rule. They have to encode (CPU) a lot of videos, which will get very few views (bandwidth).

Not really. It’s the long tail. The most popular videos are (in addition to likely being ultra resolution like vevo media’s) watched many (8-10?) orders of magnitude more times than the least popular videos. Bandwidth matters, cpu doesn’t.

It’s actually the same 80-20 rule. Eighty percent of the costs come fro serving twenty percent of their videos:)

That is a real risk - however, this time around people are already working on their own fast implementations (both software and hardware versions), so a lot of the dumb mistakes will already be avoided. There is also the risk of a lot of features being too slow to use at all in a fast encoder, but that's not a risk exclusive to AV1. HM (the reference H.265 encoder) was far too slow to be practical, and given the current speeds of H.265 encoders compared to H.264 at their best quality settings, it's not clear that they totally succeeded.

Until the bitstream is frozen, there is basically zero work on optimization.

We should abolish software patents. Reason software builds on Mathematics and you cannot patent Mathematics.

We all build on others previous work.

> We should abolish software patents.

Maybe so.

> Reason software builds on Mathematics and you cannot patent Mathematics.

This is not a convincing reason. All engineering builds on mathematics; this is not specific to software.

The argument is that software is mathematics, not build on or makes use of mathematics.

It's a small but very important distinction.

Software is not patentable in most of Europe [0]

[0] https://en.wikipedia.org/wiki/Software_patents_under_the_Eur...

But codecs are implemented in hardware. Gotcha!

Not sure if you intended this, but that is exactly the basis for the old legal workaround.

To be patentable, an invention must be "fixed in a tangible medium of expression" (or something like that). So the first software patents tried saying "we are patenting the following algorithm ... running on a hardware computer!" Amazingly they got away with it and have been getting away with it ever since.

Hm, I wonder if you could argue that running the algorithm on a virtual machine doesn't violate the patent....

Your quoted definition is for copyright, not patent. Which is why you can write a Romeo-and-Juliet story and copyright that specific story, but not the entire idea of a "doomed star-crossed lovers" story.

To be patentable in the US, an invention must be a "process, machine, manufacture, or composition of matter, or any new and useful improvement thereof". It must be at least partly new. It must be useful. And it must be non-obvious.

Abstract ideas are generally not patentable in the US, and the Supreme Court has declared that "this abstract idea, but implemented on a computer" is also generally not patentable.

I wonder if you could argue that running the algorithm on a virtual machine doesn't violate the patent...

I think Transmeta tried that by arguing that an x86 JIT doesn't infringe on patents for x86 processors.

This is great - taking a whole industry out of IP/patent litigation. The patent-revoke clause for members is converse to the idea of virality in GPL.

If I can't buy it from the iTunes store, then the fight isn't won. Once archival started becoming important to me, I've been setting about building workflows and acquiring hardware. I'm curious to see just how much of my collection will be pirated because the content industry can't be arsed to sell me a useful digital copy.

For archival use, isn't buying video from iTunes already a non-option because of DRM?

Well, yeah, that's why I would be pirating. All my music is currently in various formats, as they were bought from various outlets, my process would involve acquiring new files that don't have any restrictions on playback, hopefully all close to the same quality and bitrate.

As I'm certainly not going to repurchase it all in CD form and rip them, which is currently how the labels seem to think is the only way you should be allowed to keep such an archive, I'm just going to pirate it.

> I'm certainly not going to repurchase it all in CD form and rip them, which is currently how the labels seem to think is the only way you should be allowed to keep such an archive

In the UK the labels faught to keep this illegal. Currently, it's not lawful to rip a CD you bought and own.


I am also thinking of DRM. Thinking about it, part of the reasons music labels got big is for economy of scale when mass producing CDs for example. Of course, such economy of scale was not needed anymore with the move to digital distribution. This didn't work well with the current debt-based economy where shareholders depends on stocks always going up for things like retirements and companies treat people as "consumers" to be extracted from. I assume that Hollywood has similar problems, right?

Did they finalize AV1 by the end of the year like they said would happen? I haven't seen a performance comparison in a long time either, but it was looking good last I checked.

The Wikipedia article says January 2018: https://en.m.wikipedia.org/wiki/AOMedia_Video_1

Ah, that's unexpected. Apple decided to do a good thing for a change? Better late than never.

May be it's an indicator of general attitude change towards free codecs, and they'll now support Opus in the browser too (in OGG container). That would remove the last barrier for avoiding pointless audio storage duplication (same for video obviously).

I recently needed to choose between doing nightly cloud backups of my family's iOS photos in HEIC or JPG, and chose HEIC. Did I choose incorrectly? Which is more future proof?

Isn't it surprising that Apple is on board with AV1 so shortly after it moved iOS devices from JPG to HEIC?

I think I know what Pied Piper's crisis will be on the next season of Silicon Valley :)

People can congratulate themselves as much as they want, but the fight won't be won until VP9 is actually (technically, not legally) superior to HEVC in terms of the available implementations. Right now, VP9 encoding speeds are ludicrous and the resulting quality, as far as I know, isn't meaningfully better than what x265 produces. Having the luxury of not caring about software patents in the least, I'll stick to x265, thanks.

What are you on about?

The title is a little clickbaity, but the article is talking about AV1, it's a sort of hybrid between Daala, VP10 and Thor. Nothing to do with VP9 (well not directly).

Daala is an attempt to avoid patents by avoiding techniques used by current encoders. VP10 uses google's patents that the MPEG LA tried to fight but backed down once the DOJ got in the middle, and Thor uses Cisco's patents.

What you should get is something similar to the opus codec. A patent free state of the art codec. Preliminary tests that I know of (early 2017) showed it only slightly better that HEVC, but the goal is to have 50% better compression rate than HEVC/VP9, if they're going to achieve it or not, I have no idea.

My comment was prompted by the fact that the majority of comments in this thread relate to HEVC vs. VP9. It was my mistake to post it as a top level comment rather than a reply to one of the pertinent comments in the thread. You're right in saying that I'm in effect responding to a point that is tangential to the contents of the actual article.

Apparently no one in this thread read the article.

Is this model better? When a company develops technology with the purpose of selling the technology, it’s incentives are predictable. But what are the incentives where technology is bankrolled by companies that don’t make money from the technology, but through something else they sell? Facebook is funding these codecs because it helps their business of data mining your personal information. How is that a better state of affairs?

There is no free lunch. At least when the cost of a patent license gets bundled into your Blu-Ray player, you know what you’re paying.

You could say the same of all Web technologies. Do you actually think it would be better to have software patents and royalties on every Web standard?

For one thing, this would be catastrophic for open source projects (and indeed commercial audio and video codecs have been a huge pain for open source developers and users). Well, you're suggesting that free codecs are bad because they profit Facebook, and you could say the same about open source software, especially those that get contributions from Facebook.

I rather think that advertisers will always pay money to advertise, and when some of that money goes to developing open and free technologies that's a good thing.

There are many free lunches. You can get a universally supported free video codec to use for your hobby project or for your company's product, without having a Facebook account. It is always possible to find hidden costs in free products, so what? You can do the same for paid-for products. The idea that a Blu-Ray player won't spy on you because the manufacturer paid some royalties is... flawed.

> You could say the same of all Web technologies.

I don’t think CERN was trying to monetize anything. To the extent that web technologies today are driven by companies like Google, then maybe that’s a fair criticism too.

> Do you actually think it would be better to have software patents and royalties on every Web standard?

I don’t think it would necessarily be worse than having web technologies all developed to further the agenda of advertising companies.

> It is always possible to find hidden costs in free products, so what?

The “so what” is that hidden costs are, all else being equal, worse than transparent costs.

> The idea that a Blu-Ray player won't spy on you because the manufacturer paid some royalties is... flawed.

Paying someone up front doesn’t guarantee they won’t try to monetize you indirectly, but not paying someone up front virtually assures that they will. Compare Symbian to Android.

So on one hand a "bad" company advances its strategic interests by contributing to a free and open technology. On the other hand, this technology can bring general benefits beyond the company's interest.

The question is: does one outweigh the other? I think in many cases, for Web technologies, the corporate interest is modest and the general benefit is enormous, so it's definitely a win-win.

As for Android, the open platform part is a huge piece of tech that many have found useful beyond the smartphone market. On smartphones the open source project has enabled several ROM communities that ended up offering versions free of any Google service. There is nothing comparable on the iOS side.

Even Android with Google services is a pretty good deal: for the most part you get to choose what you share with Google (at the price of some features).

I think comparing to Symbian is difficult, Symbian never was in the same league as Android and iOS was it? A better comparison would be with Windows Phone or iOS. Apple has invested a lot in its privacy-friendly image, and I trust they are doing a good job. But they are the outlier. Then you have Microsoft... I don't know where Windows Phone standed privacy-wise, but they sure got a lot of criticism regarding privacy violations in the desktop/laptop OS. Although it's not free.

All in all I don't think Android is a good example to support your viewpoint.

>Is this model better?

I don't understand your logic, Facebook/Google is data mining your personal information with or without royalty free codecs.

The reason all these companies are banding together to create a royalty free codec is because all of them realize that online video will only become even more pervasive, and as such there is long term gain for them by getting rid of video royalties for good, and the HEVC licensing debacle must have eliminated any trace of doubt.

You can basically apply your logic to every single patent free item in the human history.

Well, having to buy a new player every year just because the codecs change is not really great so I believe open codecs are better than proprietary ones.

Even worse you can't buy a license to play the content on the device of your choice(i.e your PC). I hope Dolby will get the same medicine.

Most of the MPEG contributers aren't making their living from selling codec designs e.g. Apple, Microsoft, Nokia so many of the same concerns apply. Except with MPEG those companies have a government granted right to dictate business models to their competitors via patent licences.

It's similar to the "if you aren't paying, your the product not the customer" line, unfortunately companies can charge you for something and then also sell your data to make more money.

The trouble with most software patents is that they don't represent technology invented by companies with the purpose of selling that technology. They are made by companies doing some other business, and to put road-blocks in front of potential competitors.

Admittedly this is less true for encoding patents, some of which are held by research-oriented organisations that then charge royalties to the companies which deploy the tech.

It is better because shared resources are going into improving the technology, instead of lining the pockets of parasitic patent trolls. It should be obvious.

TL;DR: patents here slow down progress, not speed it up.

> At least when the cost of a patent license gets bundled into your Blu-Ray player, you know what you’re paying.

Rather you know what you are being ripped off for. It doesn't make it proper.

HEVC and AV1 are developed the same way: a joint effort by a consortium of companies and experts. Both cost money to make, and in both cases the consortium members expect some return on their investment. In the case of HEVC, which is a joint effort by the ITU and ISO, it’s a straightforward exchange of technology for money. How is that more “parasitic” than Facebook funding the technology to help people upload their personal data to its platform, which they then monetize through ads?

The simple exchange of goods and services for money is a really wonderful thing. Patents allow companies to do that, instead of having to give away technology and trying to monetize their users through less up front means.

> Patents allow companies to do that

Patents allow a single "blocking" patent to extract monopoly revenues over an entire area of technology. http://www.ipwatchdog.com/2017/10/11/qualcomm-antirust-war-p...

>In the case of HEVC, which is a joint effort by the ITU and ISO, it’s a straightforward exchange of technology for money.

Have you been living in a cave for last year ? HEVC licensing is an absolute mess with at least 3 different entities with which you have to settle licensing terms.

>and in both cases the consortium members expect some return on their investment.

The return on investment here is saving the cost of royalties and being in control of the development of the codec which will only become increasingly important in your business.

Now 4 different entities with the addition of Velos Media, which hasn't even announced pricing.

> How is that more “parasitic” than Facebook funding the technology to help people upload their personal data to its platform, which they then monetize through ads?

You can refuse to upload your personal data and still get the free codecs. All gain, no pain.

At some point things become more valuable as public good than they do with as something sold. Imagine if you needed to pay to use encryption like AES or ECDSA. Would the world be better off?

And yet they are slightly different:

In one case, the objective is to develop a codec that fits certain parameters, one of them being cost effective.

In the other, the objective is to cram anything patentable into the standard, in order to maximize patent fees from the users for years to come. In the past, this caused the MPEG patents to be unnecessary bloated AND having to pay for that unnecessary bloat. It was far from simple exchange of goods and services for money; it was rent seeking at its finest.

It can't be quite as simple as that. Because the amount of money they can extract from such a thing doesn't depend on the amount of patents covering it -- it only depends on how much people are willing to pay for a video codec.

What I can imagine is that every single stakeholder who had some kind of video patent would want to get a share of that pie, so they played whatever political game they could to bloat the spec to include their own patent.

It's managed very differently. HEVC and H.264 have "pools" where patent aggressors can join without actually contributing anything useful. So it's about amassing parasitic control, not about technology.

Free codecs work completely the opposite way. It's about removing the barriers of this parasitic control and allowing the codec to be used as widely as possible.

Weapons analogy describes this well. The first case is akin to arms race, where they accumulate weapons to be able to intimidate anyone when they feel like it. The second case is global disarmament.

TL;DR: it's not a technological, but rather political (control) issue.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact