AV1 and the Video Wars of 2027
hacks.mozilla.org
hacks.mozilla.org
> Tigt: I have no idea what you want me or others to do, even after reading this 3 times.
> Judy DeMocker [Author]: That’s a valid point.
I'm sure it was fun to write a post that imagines how the use of anything-other-than-AV1 ultimately leads football fans to firebomb police stations in 70 cities. But I'd rather have read a thoughtful post on actually-legitimate reasons to care about this.
I think that's actually pretty clear. Keep pushing for free and open standards. Oppose lock-in and patent protection racket.
The commenter is looking for a list of specific, actionable steps they can take to make sure AV1 "wins". Unfortunately (as the author effectively acknowledges), there really aren't any.
My worry about seeing thinkpieces like this is that Mozilla hasn't been able to influence Apple through AOMedia — it's not accidental that Apple is still missing from the front page — and so is resorting to general-audience emotional appeals. I'd go so far as to say that "here's how the sky will fall" posts like this are harmful to their cause.
In more practical down to earth terms, what you or any other regular user can do is to ditch Apple products and avoid paying them any money. Vote with your wallet.
edit - judging by the downvotes this comment is accruing it would appear Hackernews believes my cynicism is unfounded, which is quite nice.
Look, the digital video world exploded, 90% of it using MPEG's standards(MPEG1/2 to start with, h.264 now. HVEC/av1 is just a blip) they were all patent encumbered
FFMPEG runs 60% of the digital media world. _if_ patents were to be called in, places like disney, netflix & fox would have to spend literally billions swapping out FFMPEG or licensing it.
Your phone/computer pays a h264 tax, in the same way that they pay ARM/intel, qualcomm, wifi alliance, HDMI alliance et al.
HVEC gives you more efficient 4k, and HDR, but for most people that just isn't important. Streaming services still use h.264, because there is not hardware/software support upstream.
Now, the thing that will decide it is the speed/quality/cost tradeoff. If the hardware doesn't support a codec, it won't be used.
AV1 has a really slow encoder, so until that has changed it's not going to be used for on-the-fly encoding (live or just in time delivery from mezzanine codec)
for industry noise, see https://www.ibc.org/delivery/codec-wars-the-battle-between-h...
> Compared with the US, China still has plenty of room to grow in a country with a population of about 1.4 billion and 751 million internet users. The proportion of internet users who pay for video content has increased more than 10 times in four years, rising to 13.2 per cent in 2016, and is expected to further increase to 40 per cent in 2022, iResearch said. By comparison, Netflix’s US-based paying members represented around 28.1 per cent of total internet video users in the country as of the end of 2016.
https://www.scmp.com/tech/article/2137870/chinas-consumers-a...
It doesn't matter, when people or any English media speaks of the world, what they really meant was mostly English speaking World. Ignoring China most of the time.
Not the first time I have seen something like World First, World biggest etc when it is not true.
How are they not the biggest service?
Talking about paying users only, maybe Netflix is still on the top - not for long, of course.
> According to research firm IHS Markit, video streaming in China will more than quadruple from $3.5 billion in 2015, to $17.6 billion in 2020. Membership payment will take up a bigger share of the pie, increasing over 500% to $2.6 billion, with the rest coming from advertising revenue.
But talking about active users, iQiyi now have more than 500 million active users.
https://www.forbes.com/sites/ywang/2018/03/07/the-billion-do...
Yes, but it's all Chinese services combined, not just their top 1st. Unless every of those Chinese subscribers pay for the same service, Netflix is still probably the 1st worldwide paying video service.
But talking about active users, iQiyi now have more than 500 million active users.
Which is small compared to YouTube's 1.8 billion.
But fair enough, maybe YouTube and iQiyi are the biggest players; foepys was still at least half-right.
It does seem that iQiyi is using HEVC for high-res content as well, so the point still stands :)
4k on netflix is still only a small amount, as the people that have either a 4k TV, the bandwidth and device to play it on are still quite rare.
I'm all for a 'de facto' standardisation around a royalty free codec, which is what we get with AV1, but lets not pretend that these companies are doing this for any altruistic reasons.
Everything is moving to streaming, and video usage will only keep growing, so for these huge companies, having what is increasingly a core part of their business in the financial hands of third parties is an extremely unattractive position.
Google was already dead set on leaving this dependency, and now when they started work on their third generation codec (VP10) many other companies started to get the message, prompted in a large part no doubt by the HEVC patent fiasco, which now has resulted in AV1.
>AV1 has a really slow encoder, so until that has changed
Well, optimization has likely only recently started in earnest, given that the bitstream was frozen this summer. Personally I'm more interested in the RAV1e AV1 encoder in the works from Xiph than I am in the 'official' one.
Yep.
If anyone wants to know what Google and Mozilla would do in the face of corporations closing off the Internet, you can look at what they _actually did_ with Encrypted Media Extensions.
EME is bad if you believe DRM would die/be less common if no standard existed.
Firefox was forced by Chrome, and Chrome in turn was probably forced by trying to remain a viable platform instead of Netflix et al. simply providing their own desktop apps. (Though I'm only guessing, if you have direct knowledge of the what and why, please do share.)
The kind of users who are not opposed against DRM already have lots of browsers to choose between. On the other hand, people who are will now feel deeply betrayed by the Mozilla Foundation.
If you're opposed to a browser that can be used by DRM modules, Firefox was already not an option, since it supported NPAPI, and therefore DRM as Flash, Silverlight, etc. If you were using Firefox, you were already accepting that compromise. EME is simply a continuation of the status quo.
Was there an explicit or implicit covenant by Mozilla and those users?
I believe Mozilla's clear stance on DRM in the past (and its clear differentiation against other browser vendors in this respect) is near to such an implicit covenant.
Waterboarding is the better option if you believe that torture is going to happen one way or the other.
The Hollywood studios also have the power not to enforce having to use DRM on their licensees.
Tough luck with that though since EME only specifies how to communicate with DRM, and the actual DRM consists of multiple 100% proprietary black boxes that only work on blessed end devices and browsers. And since different browsers have different black boxes, you need to play ball with Apple, Google, Microsoft, and every other unique DRM vendor.
EME is a disgrace that should have never come to pass. Premium media should rather be forced to peddle their own clients entirely separate from browsers if they insist on non-cross-platform black box bullshit.
I'm not against content protection in itself, but EME was 100% marketed on false pretenses because it's just as much of a proprietary plugin system than what came before, with just as terrible cross-platform story too. If instead we had content protection schemes implemented in eg. WebAssembly, then we could actually talk about a cross-platform system that also doesn't immediately close out smaller browser vendors from the market.
The whole "DRM will happen one way or another, EME is better" is only an acceptable argument if you don't have an issue with DRM (because then you can argue that EME is "nicer" than non-EME because EME is a "standard form" of DRM).
Making it frustrating for people to access DRM content results in a distrust of DRM vendors and a general dislike of DRM -- which is (in my view) a positive outcome. Facilitating DRM in a way that makes it much easier for people to view it is not a neutral position in a world where "DRM is inevitable".
I've never heard of Youtube serving anything but h264 or VP8/VP9.
I scoured the web looking for others talking about it. I brought it up on Reddit, no one seemed to care. I still don't hear it talked about often.
Definitely felt like a "wet streets cause rain" story.
Fraunhofer essentially wrote these stories, complete with the plug for its newer (patented) audio formats so that's why they say MP3 is dead rather than hooray it's free.
If this were from a personal blog of an Mozilla employees I would have been perfectly fine with it. Instead it carries the name of Mozilla, and most of the story are simply fiction.
I don't know why the title is 2027, and 2027 was never mentioned again within the article. And for those who knows, 2027 is the year all AVC / H.264 patents expires.
I sometimes wish we could extend H.264 and create a new codec in a way, we will have a better codec that is patents free by 2027.
This is a little silly. Patent law isn’t like copyright law. The term has never changed appreciably.
It may be true that individually, each patent is mostly bullshit -- slight re-wording of prior art. But do you feel like a lucky punk, as an individual or small outfit, going up against EvilCorp's $1200/hr cap'n of their debate team in HS lawyers, a dozen times over five years? (with non-technical judges & juries as the final arbiter of your fate)
[0] https://en.wikipedia.org/wiki/Supplementary_protection_certi...
EDIT: typo
On better days I see it as a last gasp by entrenched interests to re-establish the landlordism of old now that compilers are free.
> So don’t expect that in the future you will see the progress in video compression technology that we have seen in the past 30 years.
Google has bought a whole "camp" which is badmouthing MPEG for monetary reasons. Nothing has been over-hyped as much as AV1. Who is really evil? In the end it will all be about SSIM in non-geologic time scale.
I will always sympathize with the side that does not enforce protection money (i.e. patent fees) upon others via violence (e.g. legal system).
We went from everyone backing an open codec to a dystopia with no explanation of how, only what.
The writing style is fun but makes it impossible to tell what parts are shared by our reality.
Actually, it's pretty straightforward. Everything before 2018 really happened in our reality. Everything afterward is dystopian speculation.
By difficult, I mean impossible, and there's a large number of commenters talking past eachother about this issue.
I personally am super excited for AVIF too. More revolutionary for more people than AV1 video will be? The web will make fantastic use of it in terms of page load times, and reading books in archive.org or something with instant load will be great
I've never seen 10bit h264 used outside of anime, but even there the default seem to be 8-bit h264, at least judging by what the subtitling groups use for their releases.
In terms of "put up or shut up" I should make clear I'm not advocating Mozilla shut up, more that they put up next time something like DRM in browsers is on the table (which they ordinarily do, which is why I use their browser).
In any case, don't think it will be that dystopian. There is competition from elsewhere in the world, you know.
It's encoding is so slow to the point that the encoding speed should be measured by FPD(Frames per Day)
Technical information about any of the three would be equally (ir)relevant to the article's topic: intellectual property law.
If you're really hard up for technical specs, the most cursory search might reveal https://news.ycombinator.com/item?id=16796127 regarding https://people.xiph.org/~xiphmont/demo/av1/demo1.shtml, or the part 2 follow-up https://news.ycombinator.com/item?id=17417586 regarding https://hacks.mozilla.org/2018/06/av1-next-generation-video-...
Please stop perpetuating the use of this term. There is simply no such thing as "intellectual property" and nor are there any "intellectual property laws". The term is perpetuated by proponents of the "intellectual property lobby" which hope to confuse people over what precise laws are being changed, and what precise issues are in dispute.
Not to mention that it is effectively a rewrite of history to pretend that copyright law is somehow a subset of "intellectual property law" -- the closest thing to what we would consider copyright law has existed since 1710 with the Statue of Anne in England. Laws and agreements similar to copyright (though in reality they were systems of censorship) existed even further back than that.
Copyright law, patent law, and trademark law. They are separate laws, with separate rules and regulations, separate rights, and are therefore worthy of separate discussion. Not to mention that most people (understandably) have different views on these different topics -- your opinion on whether the trademark "Coca Cola" should be used by third-parties is probably different to your opinion on whether the text of Alice in Wonderland should be used by third parties, or whether the designs for penicillin should be used by third parties.
The article in question has also (unfortunately) made the same mistake -- conflating patents and copyright law. I say "unfortunately" because I agree with the general point they're making -- that free software and patent-free standards are very important. But conflating the two doesn't make sense and just leads to confusion.
Whatever your personal objections are to a collective term for intangible human creations, claiming that "there is no such thing as intellectual property" is unhelpful to your cause, as that contention is simply absurd.
It is beyond belief that a cabal of lobbyists has convinced practically every nation on earth [1] to collectively supervise a nonexistence.
No, copyright/patent/trademark/trade secret/trade dress/design rights/moral rights/publicity rights/mask work rights/database rights/plant variety rights/geographical indications/etc(^) are not always detailed in separate laws. As early as 1916 the big three were grouped alike: see the Pan-American Convention's "Patents Trade Marks Copyright Status Report" and South Africa's "Patents, Designs, Trade Marks and Copyright Act" both in that year.
Yes, some separation exists for some of these ideas in some jurisdictions - US law separates copyright from trademark, for example, but not from moral rights, which are directly incorporated into copyright law in Title 17 USC. And no, you can't argue that copyright and moral rights are indistinguishable: even after assigning away copyright, an author retains the moral rights (e.g. right to attribution). And of course, the US also groups trade marks and patents together under the purview of the USPTO, and collectively joins invention rights (utility patents), plant variety rights, and design rights (design patents) all under the umbrella of 'patent law'.
Even where there is separation, that separation does not preclude the existence of a collective noun for the set. Nor does assignment into a collective set does imply a "rewrite of history". In fact, nearly all sets must be predated by at least one of their members; it would generally be nonsensical to create and name empty sets before any of their eventual members even exist.
In any case, arguing the merit of individual discussion is nonsequitous -- the existence of a collective name in no way precludes it.
Using your reasoning: "There simply is no such thing as the 'United States of America', nor are there any 'federal laws'. The term is perpetuated by proponents of the 'federal government'. It is a rewrite of history to pretend that that Virginia is somehow a subset of the USA -- the closest thing to what we would consider a state has existed since 1606 with the First Charter of Virginia in England. Native settlers of a similar area existed even further back than that.'
(^) Man, that's awkward. If only there was a term to denote that set that I just needed to name.
1: https://upload.wikimedia.org/wikipedia/en/a/a3/WIPO_members_...
But I found that there is actually a section on the Wikipedia article for intellectual property that outlines the basic argument against the term[1]. I understand why you might find my argument silly on its face (of course people with much more stature than me have used the term "intellectual property" before, making it silly to argue that the term itself is not used anywhere of significance) -- but my main argument is that it is not really easy to come up with a single unifying idea behind "intellectual property laws". You can fairly easily come up with a unifying idea behind "fraud laws" (use of deception to profit in some fashion), or even just "property laws" (people who have a greater claim than anyone else to something are granted certain exclusive rights to it). But the fact that even the Wikipedia article for "intellectual property" cannot describe the concept without immediately breaking it down into the specific sub-rights leads me to believe that it is fairly hard to describe the common idea underlying copyright/patent/trademark laws.
I was probably too forceful in my original response, which wasn't really fitting (since I actually agreed with the general point being made).
> (^) Man, that's awkward. If only there was a term to denote that set that I just needed to name.
I mean, my point is that the only time you need to name them together is when discussing "intellectual property". When it actually comes to filing a patent, or registering a trademark, or suing over a copyright infringement claim, you never would reference those laws together. I agree it's awkward, but I disagree that you would ever need to list them like that (because they are naturally separate concepts that aren't much more related than any other legal concept).
[1]: https://en.wikipedia.org/wiki/Intellectual_property#The_term...
Examine https://en.wikipedia.org/wiki/Admiralty_law and see that in the opening summary the article similarly resorts to naming "marine commerce, marine navigation, salvage, maritime pollution, seafarers’ rights, and the carriage by sea of both passengers and goods" while distinguishing it from the Law of the Sea -- which is again immediately defined-by-list: "navigational rights, mineral rights, jurisdiction over coastal waters, and the maritime relationships between nations."
Sadly, no, "most people" don't feel differently about these rights. In fact, most people cannot even differentiate one from another. This is certainly made no better by the growing use of the umbrella term, but is not caused by it. Search any online art site and find thousands of artists disclaiming "<character> is copyright <entity>", when in fact the only copyright in question is held by the author himself on the drawing he just drew. The artist intends to forswear the trademark, but cannot or at least does not distinguish these legal ideas.
EDIT:
Removed further arguments, because in actuality I agree with you (and this article). My original reaction was a kneejerk to your admittedly overstated claim of nonexistence.
I too am concerned that greed-motivated lobbyists deliberately distort issues to strengthen intellectual property laws against the public interest. Your linked wikipedia section quotes Stallman and Lessig, as well as economists who would prefer the more descriptive term "intellectual monopoly", all objecting to to property's false implication of scarcity.
> Your linked wikipedia section quotes Stallman and Lessig, as well as economists who would prefer the more descriptive term "intellectual monopoly", all objecting to to property's false implication of scarcity.
Stallman does actually believe that the term "intellectual property" to group copyright/patent/trademark/... laws is intentionally deceptive and causes confusion[1]. Personally I don't quite agree with all of his arguments, but he does have a far more radical view than the others mentioned in the article.