> DRM is a hugely important part of the online video workflow and HTML5 offers a very useful tool with Encrypted Media Extensions.
DRM has never been important for online video workflow and it's only useful in the way a walled garden is useful... it will require that each pair of browser-website (with EME) negotiate before allowing it, basically making it impossible for smaller browsers to become useful at all
Flash had video long before HTML5 did, and HTML5 video support has only very recently been consistent (e.g., the codec wars). Those made Flash a better option for a long time, regardless of DRM, and might also be reasons why Flash has stuck around in this space. DRM is not the whole story.
No it hasn't. It's just been a demand from media execs that distribution networks are forced to implement.
Netflix is competing with free and winning -- they've created something that's actually more convenient than piracy. I can already download everything on Netflix from my friendly local piracy site, but I don't because the eight bucks a month is worth less than the time I would waste pirating it.
Hell, Netflix could put a download button next to all their content and I bet it wouldn't affect piracy. The value isn't in the content but the distribution network that's so good even pirates can't compete with it.
Meh. Netflix would not affect piracy because their offer is very small. Netflix is good if you're an American and not very into movies. As soon as you are looking for something outside the "very mainstream" you're out of luck.
Your "I can already download everything on Netflix from my friendly local piracy site" works, but the other way around doesn't. I've been a Netflix user for years, but I'm on the verge of unsubscribing, because I realized that since January I've only watched House of Cards.
I can also offer anecdotal evidence that the piracy rates of my college social network dramatically decreased as the affordability and content quality of Netflix increased.
Netflix is being squeezed by media publishers which is why they started publishing their own content. I still hope that streaming services like Netflix will win and reshape the market...but it is a difficult fight.
On the note of you only having watched "House of Cards", there are a bunch of Netflix shows I could recommend for you if you're interested. "Peaky Blinders" comes to mind immediately since we are due for another season at the end of this month.
Absolutely. The only case where (I prefer to call it "morally ambiguous streaming") wins is in series with ongoing seasons. But Netflix is so fast and convenient, I usually end up waiting for it to appear on Netflix, or rewatching it there.
Would there be more piracy of online video had no DRM? It only takes one person to try hard enough to rip a video and put it on a torrent. Conversely, I don't think removing DRM would increase viewers (not that many would not watch a movie they want to see due to their thoughts about DRM), but at least the providers could save a bit of money without it.
I do think removing DRM would increase viewers in the long term: the data has always been the valuable part, and opening it allows innovation in interface, suggestions, and integrations.
Unfortunately, it seems like netflix's desire to open their catalogue at all decreases year after year. Their competitors are even worse in this respect.
> Conversely, I don't think removing DRM would increase viewers (not that many would not watch a movie they want to see due to their thoughts about DRM)
My TV provider recently "upgraded" the interface on their set top boxes to one that is so slow and poorly laid out that I've just stopped watching two premium channel shows because it now takes five minutes to navigate to the point that I can watch anything, and am going to stop subscribing to those channels. This is DRM costing them viewers and money.
There's a difference between "In a perfect world, DRM is necessary" and "In the world we live in, DRM is useful".
To make a possibly-emotionally-charged analogy, it's like patent licenses for free software. In a perfect world, software patents wouldn't exist. But in the world we live in now, you cannot build practical systems without acknowledging that they do exist and they need to be dealt with, and that you need to be as clear about patent rights as you are about copyright. So e.g. the Apache License and GPLv3 have some text about software patents and how they interact. It is hugely important for some people (depending on their jurisdiction, quality of counsel, amount they care, etc.) that they use licenses like those that acknowledge software patents. That's not to say that software patents are inherently useful, and certainly we should work towards a world where they're not, but if you're building a real project, you have to work with the world as it is.
In the world we live in, the people who own the content that lots of people want to see have decided that they care about DRM. Whether they should care about DRM (or whether the viewers should want to view that content, for that matter) is beside the point for the discussion of whether DRM is currently useful. By all means—including the means of making DRM technically cumbersome—convince them to stop caring about DRM. But the ability to provide DRM is, unfortunately, useful today.
> But the ability to provide DRM is, unfortunately, useful today.
It's useful in the sense that snake oil is useful after somebody signs a contract to use snake oil. That doesn't mean it's actually useful, it means somebody signed a dumb contract.
When that happens you can go out and take bids from snake oil suppliers and rub snake oil on your skin every day that does nothing but cost money and make you smell bad, or you can fix the contract.
With what? EME is a "cure" for a problem you shouldn't have in the first place. Idiotic technology to do nothing useful and solve no problems to make some content company executives feel better about their liability.
EME does abosolutely nothing to prevent piracy and only serves as another roadblock for people to jump before seeing videos.
All it shows is that there is certain group of people which will pirate even if the content would be given for free.
The people that would view paid content should not be punished with DRM, that only works on certain platforms and certain browsers.
It actually makes pirated product much better than original. Pirated videos besides being free work on all platforms, you can view it on your computer, phone, tablet. Living or visiting a different country does not revoke your rights to see it, and you can make as many copies of it as you want.
Well, if youtube clips uploaded at 1080p couldn't be watched at 1080p, people would complain. When services advertise "hd video" and it doesn't actually work in HD... I'm sure you can figure it out. But then maybe you can't, given your remark.
The most attractive feature of netflix is that it is everywhere and most people would not have a problem installing a dedicated app, just look and android and ios.
Netflix already has their own Windows desktop app. Because it can use Microsoft's unsandboxed PlayReady DRM, it can play 1080p video while Chrome and Firefox can only play 720p.
What bugs me is that Google gave itself a pass by white-listing YouTube - because it turns out some people still need Flash on YouTube. Google should dog-food their own policies.
Likewise, Google penalizes webpages by size. A 3 megabyte page gets a larger penalty than a 1 megabyte, which has a larger penalty than a 1 kB page.
But, they give themselves a free pass, as their own "webpage guidelines" page is 3.4 megabytes long! But, even worse, it has a video on a carousel that repeatedly downloads the video whenever it restarts the video loop.
How do you know they give their own web page a pass? It could be scored low and Google just doesn't care, or like all of us, has other priorities. The fact that Google recommends lightweight pages, and still builds heavy pages is contradicting, but doesn't mean they've built in some sort of "whitelist" for their own sites.
Google doesn't make websites be lightweight. Heavy websites get indexed just like any other. If they want to make a heavy page and accept the penalty, there's nothing wrong with that.
Are there parts of YouTube that require Flash in Chrome other than legacy embeds on non-YouTube sites that enable the embedding page to script the Flash embed via a JS API?
>other than legacy embeds on non-YouTube sites that enable the embedding page to script the Flash embed via a JS API
That's probably it. Then again, it's not just Google that has to deal with legacy applications. So yeah, if you're going to shove this policy in the name of open standards, dog-food it and deal with the same pain countless of companies will have to.
Flash is only required when subscribed to Google Play Music All Access / Youtube Red, otherwise a free account can readily enable HTML5 audio in the options, which works in Chrome/Safari/Firefox (just tested again to be sure) and probably Edge also.
Google Play Music uses MSE (Media Source Extensions) to stream MP3. Firefox doesn't support MP3 in MSE streams. Google Play Music could use AAC or Opus in MSE, like YouTube does, but they prefer MP3 because some Android devices have MP3 hardware decoders that (may!) use less battery power than software decoders.
YouTube encodes all their content with in multiple formats: AAC and Opus, H.264 and VP9. I don't know why Google Play Music can't encode their music streams in both MP3 and Opus. They would save on bandwidth by using Opus.
which i use in a lot of projects that need simple charting. i really don't know why google doesn't dogfood this product more, its a great, simple, easy to use charting system.
EME is a huge advantage from a security perspective because it avoids pulling in a massive platform (Flash, Silverlight) just to play video. We can debate other aspects but from a security perspective moving to a small, tightly-scoped module is a big win and that's important when the cost of vulnerabilities is in the billions per year.
It's also somewhat good for the web since both of those platforms are competitors and companies often reuse things they've already invested in.
It is unequivocally bad that it poses a risk to Mozilla but the real problem isn't EME but rather the fact that all of the DRM opposition since the 90s has failed to move public opinion much. As long as customers happily pay for DRMed content and three of the major browsers are made by DRM vendors, the most likely alternative to Flash is a proprietary interface. The fact that EME is standardized at least gives the EFF better grounds for demanding consistent treatment.
You don't need EME to play video, HTML5 video can be played just fine without it.
AFAIK, You need EME to make sure that only browsers that have the Foo-Corporation EME plugin can play Foo-Corporation videos.
And to get the EME plugin for a browser, you need to ask permissions from Foo-Corporation to have it. So anyone writing a new browers (like Mozilla did with what became Firefox) will have to ask for this permission and will likely not just get it.
Seems like another example of where big companies got to where they are because the space they moved into was wide and open and free. But now that they have claimed that space and divided it up amongst a few players, they'll do their damned best to fence it of, wall it in and make sure none else can ever move into ever again.
> You don't need EME to play video, HTML5 video can be played just fine without it.
I'm quite aware of that but consider it from the perspective of a normal user using a non-Apple/Google/Microsoft browser:
1. Click on a link on YouTube, Hulu, etc. It's not DRMed, so it just works.
2. The next thing in the playlist is DRMed (e.g. it's premium, owned by a more restrictive publisher, etc.) and it doesn't play.
For people like us, step 3 might be “Decide you didn't want to play it that much” or “Complain to the provider”.
For the vast majority of users, step 3 is more like “Give up on Firefox and use Chrome/Edge/Safari”. How many of them will eventually stop using Firefox in the first place?
That's the real problem here – even if Firefox had 90% marketshare, it would evaporate quickly because switching is easy, users have a bunch of reasonable quality options, and most users don't feel that they're getting a bad deal.
EDIT: note that I'm not saying this is how I'd like it to work, only that anyone who thinks DRM is unjust needs to think about the millions of Netflix subscribers and both why and how those people should spend their $9/month elsewhere.
But the vast majority of video watched today (think Netflix) would not be on the web if it were not DRMed.
Philosophical arguments about what "video" "needs" are all very well, but the side that actually provides users with the content they want where they want it will be the side that wins, and very quickly.
Leaving aside the ethical stuff what worries me most about EME is that it'll break my linux usage at as a desktop, with a few exceptions I currently have a web experience equal to (or better) than anything on Windows or MacOSX, binary blobs that may never make it to Linux are a concern.
Agreed — I don't see much chance for change in the general DRM debate but it would be nice if it at least hit the point of something like H.264 where there were standard terms available to anyone who wanted to implement it so a new player couldn't be frozen out.
(Back in the day, we had a client who was working on a BeOS appliance. Plugins were a big deal since things like Flash were theoretically portable but Adobe wasn't going to talk with you for less than a certain dollar amount)
The only problem I have with supporting EME is that it closes the market for new browsers to enter because they need to get certification. What incentive do the current market players have for allowing new browsers access to EME?
> but Encrypted Media Extensions are definitly not an advantage
Yes, it is dishonest of Google to keep referring to 'the Open Web' when 'HTML5 video' is heading towards being simply an enabler for other proprietary plug-ins.
No, but they do need to get a signed chain-of-trust certificate rooted to a certificate authority recognized by their clients' user agents. They also need a well-defined domain name that was registered through some domain-name registrar, somewhere.
The notion that some parts of the web standard are pay-to-play is actually not new.
This is a real blow to sites that rely on Flash's RTMP for low-latency live streaming. There's no real HTML5 equivalent right now. HLS and MPEG-DASH have a delay of around 30 seconds, often more.
Though not imposed by the technology most of your live TV isn't actually live. There's a thing known as the broadcast delay or the seven seconds delay that a lot of broadcasters use to beep out profanities or do live subtitles and other alterations.
I did a cursory search out of curiosity: TV has a 7 second delay they add themselves to filter content. The broadcast itself is pretty much instant. It's impressive.
I work for a church that does live broadcasts. We see about a 3-5 second delay on OTA and about a 4-6 second delay on cable. We are tied to the tv station by a fiber line and they are only a block or so away.
Interestingly if you put a TV using cable next to a TV using satellite the satellite TV will trail cable by 2-3 seconds sometimes just from the amount of time to bounce the signal off of satellites.
Yes, we depend on Flash for this, too. I worry about Flash disappearing from the ecosystem before HTML5 streaming matures enough to deliver a decent, cross-platform-consistent live-streaming experience.
Including Google's own Google Music, I use Firefox in my daily browsing without Flash installed, when I want to listen to my music i've to start Google Chrome which has Flash by default. The experimental HTML5-audio toggle is always disabled for me.
Play Music always uses HTML5 now, which is why the toggle is not available (but it's indicated badly). You can disable the built-in Flash on chrome://plugins and everything will still work.
Only in Chrome, it still requires Flash in Firefox for some reason. They're using some non-standard way of doing HTML5 Audio as far as I remember, which only works in Chrome.
Bug report is here - https://bugzilla.mozilla.org/show_bug.cgi?id=911837 - It's been a while since I looked into it and looks like the details have changed recently. At least a quick attempt now tells me "Please refresh this page in your browser to get back to listening to your music. If issues persist, make sure you have Flash enabled and working."
Google Play Music uses MSE (Media Source Extensions) to stream MP3. Firefox doesn't support MP3 in MSE streams. Google Play Music could use AAC or Opus in MSE, like YouTube does, but they prefer MP3 because some Android devices have MP3 hardware decoders that (may!) use less battery power than software decoders.
YouTube encodes all their content with in multiple formats: AAC and Opus, H.264 and VP9. I don't know why Google Play Music can't encode their music streams in both MP3 and Opus. They would save on bandwidth by using Opus.
That's like banning petrol cars and forcing everyone to go electric (because it's possible) or walk! Why can't kids enjoy some games after all? Maybe the companies behind don't have the funds to rebuild their free-to-play games just because Google doesn't like Flash hence something valuable is lost
It's not as simple as "Google doesn't like Flash."
Flash is a known attack vector. I'd be more comfortable with the argument "Kids should be free to play games" if I didn't have to keep purging virii off my aunt's computer because she won't stop visiting Flash game sites.
Besides, there are alternative browsers. If kids need to drop into Edge to play Flash games, that seems like an acceptable solution? If it isn't possible (because they're on a school network that whitelists browsers, for example), that's indicative of the cost / benefit tradeoff of that use case.
There are alternative browsers indeed and probably some will migrate but many won't and the Flash platform will suffer. Chrome, being dominant on the market, has the power to kill anything that doesn t align with the company s interests, which means we don t have an open web.
This article has the vibe of drawing attention to something else other than the subject of the piece. Is that unfair to say? Not for me to judge I suppose, but the final paragraph says it all.
Anyway, a couple of things from the article...
> "Netflix have already switched to HTML5...due to the benefits of HTML5 based streaming"
> "...HTML5 outperforms plugin based playback in almost every department"
> "...delivering high bandwidth products such as 360° video and Virtual Reality in a more efficient manor."
Such generic statements! Netflix HTML5 playback in both Chrome and Firefox is limited to 720p. That's not a benefit and is why I must use silverlight in IE on Windows 7 to watch 1080p Netflix. They really should cover things like this to better reflect how things actually are.
The claims about HTML5 performance over plugins should only come from people who have done side by side comparisons. So many bloggers telling it according to something they read somewhere.
Finally, "360° video in a more efficient manor". I like efficient manors, even bad manors, but again this claim should be backed up with evidence. If you look at VR, if you look at 360 video, the evidence for better performance in HTML5 for these technologies is simply not there. Even with still image panoramas, anyone who has made these in something like KrPano will know that Flash panoramas generally outperform HTML5 panoramas. Smoother motion, better memory management, better cross-browser performance. We may wish it weren't so, but we should get facts straight regardless of whether it leaves HTML (which we love) a little bruised in the comparison.
While the focus is rightly on video, there are a huge number of games written in flash. In the short term this s probably a hug boost for unity... But as flash dies there will be a lot of games that die with it.
Not to mention the massive elearning industry which has many, many years' worth of training content in Flash. Products like Captivate and Articulate can export to HTML5 these days, but converting existing content is a huge feat.
Unity is itself a proprietary, single-vendor, closed-source plugin.
Though I don't disagree at all, transitioning from one proprietary, single-vendor, closed-source plugin to another isn't actually going to solve the problems that killing Flash is supposed to solve (apart from the one where Flash is also a gigantic legacy-code monument to all our sins, of course ;) ).
From what I remember it's not as performant or reliable as the plugin, so there's some work to be done on unity's end. Beyond that, if browser makers want to replace native code plugins they have to give a better alternative. WebAssembly might help, but since it's essentially running in the same vm as JS I'm not sure it'll be faster other than the parsing overhead.
> From what I remember it's not as performant or reliable as the plugin...
... two things that absolutely kill videogames, so it's more accurate to say "Unity you will output WebGl-based games" (at some indeterminate point in the future).
Agreed that it's a step in the right direction over alternatives though.
Chrome already killed the unity plugin which is what really lit a fire under there ass to get WebGL working well. Its reasonable now, at least for some types of games.
Getting really frustrated with all the things chrome is starting to change.
We are starting to become a bunch of sheep, if google wants something, they announce it, and implement it, we developers, users have to fall in line and start doing everything they want.
What the hell is going on, why are we not pissed off about this. Look I hate Flash as much as the other guy, but blocking an extention just because a company doesn't like it. How long before we see adblockers, anti piracy trackers and basically anything that a company like google will see as a threat get blocked.
Don't give me that "Oh flash has security holes so we are blocking it." Yeh when our governments use this same excuse we are all up in arms, but when a company does this many are okay with it.
The sad sad truth is majority of users are going to fall in line whenever a company decides what's best for them.
</rant>
And to all the fan-boys/girls down-voting me. Have the decency to leave a comment as to why you are downvoting.
I'll see your rant and raise you an exasperated sigh.
> We are starting to become a bunch of sheep
You might be, but not all of us are.
> why are we not pissed off about this
About them blocking (but still bundling) flash in 2016? I'd honestly be more questioning why they ever bundled it in the first place. I use Safari and have not installed Flash in.. years.
But if you mean pissed off about Google using it's position of power to arbitrarily decide what technology "wins" and "loses" - people are pissed off. Unfortunately, there is seemingly an entire generation of developers who are somehow perpetually convinced Google is the second coming of jesus, and that anything they do must be a good thing.
I've seen developers who will argue that they use emacs/vim over a full IDE "because now I'm not reliant, I can work on a server if i need to" and then use Google for fucking everything. So the lesson is, you can't become dependent on something you actually have control over, like a laptop, but it's fine to become dependent on a mega-corporation that has a history of just abandoning shit because nothing ever lives up to the ridiculous amounts of profits made by being a creepy personal data collecting tracker of fucking everything.
Regardless of your thoughts about flash, this is the perfect opportunity to explore the world outside google. Other browsers exist. Other search engines exist. Other ad networks (and less fucking creepy ones) exist. Go forth. Be adventurous. Be free. Seriously.
Anyone who thinks Google's level of control on the web (search engine, browser, data centres, home internet connections, mobile phones) isn't a massive risk is kidding themselves.
Edit: for clarity, I'm not arguing that the situation isn't fucked, just saying, don't lie down and accept it. Also, have an up vote, because, you know, the voting system here is fucked.
> About them blocking (but still bundling) flash in 2016? I'd honestly be more questioning why they ever bundled it in the first place.
Google bundles Flash with Chrome so they can push out Flash security updates quickly and seamlessly.
Adobe's non-Chrome Flash Player has its own update check, but it's not automatic. The "updater" just opens the adobe.com download page. The user must still manually download the real updater exe, unselect any "special offers", run the updater from wherever their browser saved it, and restart their non-Chrome browser.
I wonder how common Flash would still be if Google hadn't bundled Flash (and paid Google security engineers fuzz and fix Flash bugs).
> But if you mean pissed off about Google using it's position of power to arbitrarily decide what technology "wins" and "loses" - people are pissed off.
I think categorizing it as winners and losers is wrong. As much ribbing as google gets, I think very often, they are using their position of power to make the internet more secure and functional. They won't get it right every time, but I'm happy to see them try.
Flash has other problems besides security exploits.
* It's a single-vendor/proprietary runtime.
* It's not even available for mobile. Most of the content that uses it is made for desktops.
* It's a black box.
* It doesn't work great on platforms Adobe doesn't prioritize (if it's even available).
* We have to wait for Adobe to fix bugs and vulnerabilities.
All Chrome is doing here is phasing out the Flash Player runtime.
Even when support is dropped entirely you can still use something like OpenFL and/or Haxe and target the web with it if you still want to use your Flash skills.
You're arguing for the tyranny of the statuo quo. It doesn't make sense for Flash to even exist, it's only around for historical reasons.
If we all had browsers without Flash and someone came along and suggested adding it, what would your argument be? I don't see why our default should be not changing things when everyone knows the better technical answer is changing it.
That is true. But they are in a position of incredible power.
They have control over a lot of what people see online, by virtue of controlling what people find when they search. They're so ingrained in society that they're a verb. Their staff have had remarkable access to the White House and other government institutions.
No, not a monopoly...but they're not far from potentially becoming one.
The existence of alternative browsers. If your company is relying on Flash, Chrome just put itself out of the running for being your enterprise's solution. By all means, use the browsers you need to get your work done.
The fact this is a price Chrome is willing to pay indicates just how dead Flash is, and how much browser vendors are willing to kill it.
You seem to live in the bubble where everything in the world is written using the latest technology at your disposal.
The sad reality is, there are are many companies that depend on flash for their corporate software. We all hear horror stories where IE6 has to be supported.
While I whole heartedly agree flash is something I would avoid like the plague it is not the reality for many companies. Think about it, at one point we all though flash was the cool new thing, but now almost everyone we know hates it. But if you force thousands of companies to replace their software just becuase another company does not want to support an ancient technology, it should ring alarm bells for us all. It would cost billions of dollars in upgrades if we consider all the companies that would be effected.
Google has to be put on the spotlight when they make changes such as these that would effect many people they have to give a good reason other-than, "oh its old, has many security bugs."
Google has to stop forcing everyone to conform to what they think is a best practice. They need to go through the proper channels, they need to get people like the etf, w3c and take these actions in a democratic way.
Google is slowly moving away from what made many of us love them, "Don't be evil" google was our hero in a time we had to put up with companies like Microsoft. It makes me incredibly mad to see a company I loved turn in to the companies they replaced.
I do hear horror stories where IE6 has to be supported.
The solution to those in the aggregate is not "Maintain IE6 backwards-compatibility forever." The challenge associated with that is why they're horror stories. But that's the opposite of what you're describing here; IE6 being supported ad-nauseum was necessary because MS had a bunch of non-compatible proprietary features in IE6 that didn't match to web standards. Google is trying to move out of supporting a non-web-standards architecture.
I'm confused about what aspect of minimizing the attack vector on a standard user's computer fails to meet "Don't be evil" as a motto. Is it the part where minimizing the attack vector for people who don't have a need for Flash collides badly with the needs of people who do have a need for Flash? People who still have a need for Flash have alternate user agents available to them; they don't need Chrome.
I'm sorry this change means you're going to have to account for your enterprise's reliance on Flash. I have friends who had to maintain code backwards-compatibility to pre-generics versions of Java. Enterprises gonna enterprise; we should neither hang the cutting edge nor the common person use-case on the relative lethargy of large companies.
In a world where we're ranting about Flash (a non-standard, closed-source web plug-in with a few ad-hoc, reverse-engineered, quite-possibly-illegal alternative implementations) being buried and replaced by a web standard---even a web standard primarily driven by a handful of web development incumbents, but a standard nonetheless---we are truly through the looking glass.
What do you mean by “blocked”? Chrome on Android has never supported
any sort of extension, I wouldn’t say that they’re blocking
ad-blockers, more there’s just no support for ad-blockers, or
extensions for that matter.
I've heard that many CDNs are still lagging to adopt DASH, and it causes a major problem for any service which wants to implement HTML5 video (DASH + Media Source Extensions). How does it work for Youtube, do they simply run their own CDN?
> DRM is a hugely important part of the online video workflow
I agree to other commenters here who pointed the nonsense of this. DRM is not an important part of the workflow. It's a tool of corrupted groups for standards poisoning, market control and "creative" undemocratic lawmaking.
Maybe I'm naïve or not seeing something here but DASH looks like something you can do by storing a couple of files somewhere and setting a MIME type [0]?
What's special about DASH that CDNs need to specially support it? Or are you talking about services that handle transcoding video and such as well?
From what I understood, CDN is needed to lift the heavy load from the source servers. Without them you'd be hammered by heavy traffic, and not only that, the route to the end user will be most likely longer.
DASH is adaptive streaming, and as such, when you select some stream from the client, CDN should support that in their infrastructure (including codecs and etc.). I don't know the details exactly, but it's far from trivial. May be someone can explain what the difficulty is in detail.
Currently CDNs support HLS and Adobe's adaptive streaming, both of which aren't open standards and aren't part of HTML5 video.
Without DASH support in CDN, you can't scale your service, and to get rid of Flash and Adobe's adaptive streaming you need DASH + MSE. That's how I understood it at least.
> CDN should support that in their infrastructure (including codecs and etc.)
This makes sense if the CDN is transcoding the video for you but if you do that yourself I don't think a special CDN is necessary.
From Mozilla [0]:
> DASH works via HTTP, so as long as your HTTP server supports byte range requests, and it's set up to serve .mpd files with mimetype="application/dash+xml", then you're all set.
And from encoding.com [1]:
> To play the content, the DASH client first obtains the MPD. The MPD can be delivered using HTTP [...]
> Using this information, the DASH client selects the appropriate encoded alternative and starts streaming the content by fetching the segments using HTTP GET requests.
This definitely sounds like it doesn't need any special support from the server. As long as it can serve files over HTTP, the only special thing is the transcoding and setting up the manifest, which doesn't need to be done by the CDN.
I have a feeling CDNs advertise support for HLS and Adobe's streaming because they require additional support.
> DASH is adaptive streaming, and as such, when you select some stream from the client, CDN should support that in their infrastructure
Adaptive streaming is essentially a client side technology, all the CDN needs to do is ensure that it serves up files with the right mime type (not actually a hard requirement, but many players choke if you don't).
Which segment to play next is decided by the player based on current playback conditions and a manifest of available segments and bitrates; it then requests the relevant segment from the CDN, which delivers it as it would a normal file. (A segment is just a very short video file).
The CDN is not expected to keep state for each connected client, which it would need to in order to be able to make the relevant decisions on which file to serve.
For live, you also have the complication that you need to keep adding files, and updating the manifest to indicate where these new files are (The payer checks back every couple of seconds to see if the manifest has changed). Any latency at all in this means that a client may get back a stale manifest, and then not know where the next video segment is supposed to come from, causing stalling. (Numerous ways to work around this, but it is the real hurdle to easy adaptive live streaming).
This is all a gross simplification, but at its most base level, that is how it works.
You can (theoretically) do MPEG-DASH from a normal MP4 file, but that does require CDN support (and possibly a single pass to add more info at segment boundaries).
> Currently CDNs support HLS and Adobe's adaptive streaming, both of which aren't open standards and aren't part of HTML5 video.
HLS is an open standard[0], though not defined by a standards body; more Apple have documented how they do it and others have adopted it as a defacto standard.
And that is the biggest hurdle for DASH adoption - Apple devices.
As a content provider, I have* to support HLS if I want to support adaptive streaming in iOS. Transcoding to another standard (one which is still in its infancy) is just asking for non essential costs and headaches.
*Yes, it is possible to use MPEG-DASH, but not natively supported by Apple, so you are on your own.
Until Apple support MPEG-DASH you still need to provide a HLS stream.
DASH is a sprawling standard, and in theory you can use DASH with existing HLS segments (MPEG-2 .ts files), but no-one supports it (or likely ever will). DASH264 [1] is an attempt to keep it simple, and tie it down to a smaller subset of codecs (x264/5 and the like), and seems to be making headway.
It's early days for MPEG-DASH, and it is gaining traction, helped by the fact it has been learnt a lot of lessons from the likes of Flash and HLS.
Twitch definitely offers HLS as well; it makes sense to support the current generation of devices, but re-use the tech stack for future standards. Clever peeps.
> As a content provider, I have to support HLS if I want to support adaptive streaming in iOS.
I see. Safari on iOS is still missing MSE? Is Apple doing it to slow down DASH adoption?
The other expected problem is codecs. Apple are being sleazy and refuse to support free codecs like VP9 and Opus. So making DASH serving those won't work with Apple too (unless you'll use pure JavaScript decoder which is very not optimal).
May be Daala / Thor will eventually break their nasty behavior, especially if it will be adopted as mandatory for WebRTC.
> DASH264 is an attempt to keep it simple, and tie it down to a smaller subset of codecs (x264/5 and the like), and seems to be making headway.
Tying DASH to a set of proprietary codecs to propose a standard sounds like a very bad idea. I don't see it going anywhere.
> Apple are being sleazy and refuse to support free codecs like VP9 and Opus.
> May be Daala / Thor will eventually break their nasty behavior, especially if it will be adopted as mandatory for WebRTC.
I think AV1 (Alliance for Open Media's codec) is likely to be the video codec Apple joins in on, unless it tries to push HEVC. With Intel, AMD and ARM in the group, basically every computing device is going to have hardware support for it so Apple would have to actively cripple it in their hardware to go without.
As for audio, Opus is great for low latency and low bitrate but AAC is widely available and supported and performs better than Opus at high bitrates (which I suspect are more important to Apple than low bitrates). Maybe if Opus becomes part of a telecom spec for VoIP or something Apple will pick it up but I think it's unlikely otherwise.
AV1 is the new codename for Daala + Thor? They used to call it NetVC. I hope they'll come up with better name than something that sounds like a virus :)
> As for audio, Opus is great for low latency and low bitrate but AAC is widely available and supported and performs better than Opus at high bitrates
Opus beats AAC even for music from what I saw and provides same quality at lower bitrates. I.e. it has lower transparency level, which means it's simply better overall. See http://listening-test.coresv.net/results.htm
No, AV1 is the name of the codec the Alliance for Open Media is producing. The Alliance for Open Media is a coalition formed by Amazon, AMD, ARM, Cisco, Google, Intel, Ittiam, Microsoft, Mozilla, Netflix and NVIDIA [0]. I don't believe Daala is involved.
> Opus beats AAC even for music from what I saw and provides same quality at lower bitrates.
That test notably omits the proprietary Fraunhofer AAC codec, which may outperform the others.
It also produces results differing from the SoundExpert ratings [1], which show Opus and AAC both achieving transparency around 128K, but with AAC doing it by a more comfortable margin.
> Plus AAC is not free.
Freedom isn't relevant to what will be adopted. The parties in control aren't motivated by freedom.
> Freedom isn't relevant to what will be adopted. The parties in control aren't motivated by freedom.
Oh, they are. It's only the matter of scale and who owns the non-free options. If those who have adopt some codecs need to pay tons of money for non-free ones they start caring about
freedom quite a bit ;)
That's exactly why HEVC shot itself in the foot and will eventually die miserably.
And how is it supposed to be used if CDNs don't support it? Sounds like a circular dependency to me. Technology is here, but without infrastructure it's close to impossible to use at scale.
Catch 22. To cut Flash out you need to use DASH + MSE. You come to CDN and say you need it, and they say they don't support it because you first need to use it...
Google worked around it by rolling out their own CDN. Do you propose to do the same?
Because the internet is no longer fair :( Things are breaking, I cannot get a mail server set up to send mail to microsoft owned properties, no bounce message just a void, I'm 10/10 on mail-test .com, DMARC, SPF, PTR, no RBL, just m$ don't want me to send email to their users, a private RBL maybe, I cannot know because they don't reply to communications regarding it, nothing to do with this point, except it's asymptotic of a capitalist system with no custodians, nobody is being paid to keep it fair, so it won't stay fair, education of the public is a possible hope, but again, nobody pays for that so...
Could this be IPv6-related? I've run my own server for a long time, but when I spun up a new VPS and enabled IPv6, I suddenly couldn't send mail to Comcast customers. It seems that the single default IPv6 address fell into a privately-blacklisted /64. I was able to work around it by firewalling Comcast's IPv6 MX hosts, but getting my own /64 was the real fix.
On the other hand, I don't know anyone with an @some-microsoft-property email address, so I don't know whether I'll be having trouble sending things there. I can reach Gmail and Yahoo (and after getting my own /64, Comcast) without trouble, though.
EDIT: Hmmm, mail-test.com seems to be a parked domain.
A college of my has struggled days with this issue, gmail wouldn't accept any mail send from a IPv6, ended up disabling IPv6 at the end. It seems like google only uses IPv6 for internal mail's, sending a mail to another account sends them using IPv4. Even Facebook, Microsoft etc still use IPv4 for their email's.
Google do of course use IPv6 externally, but require rDNS for all inbound IPv6 SMTP sessions. Without that set for your MTA's host they'll reject immediately.
thanks for the thought, I'll check but pretty sure I', stil on IPv4, 6 not enabled, maybe I'll try enabling it, or burning some sort of effigy, http://www.mail-tester.com/ is the link I intended.
Okay, jumped some hoops for m$ to try and get my email working, but no joy, sender.office.com reports the IP is not blocked and suggests I refer to bounce error (there is no bounced message), I tried the Outlook mitigation route, and after 24 hours received a "Not qualified for mitigation
... Our investigation has determined that the above IP(s) do not qualify for mitigation." message (my stars). Today I've joined the Microsoft Smart Network Data service, and that reports the IP of the mailserver as having normal service, so where now ? PM Bill?
They do maintain their own blacklist, which doesn't seem to be listed on any of the public monitors. When I had this issue it did give a helpful bounce message:
Client host blocked using
FBLW15; To request removal from this list please forward this message to
delist@messaging.microsoft.com (in reply to RCPT TO command)
I did as it said and forwarded the message to that address, and got an automated reply and my IP removed a few days later.
Maybe your situation is different since you're not getting a bounce message, but try emailing that address if you haven't already.
Nothing Google does can surprise me while I can still say 'OK Google' to 'OK Google' and it think I want to search for 'OK Google', who ever searches for 'OK Google' ? Or maybe it's the Android N preview that bricked my phone that explains my mood.
Because Google of 2010s is the same as Microsoft of the 1990s. The only difference being most of the tech world in the 90s realized how badly Microsoft was abusing their position and almost nobody complains about Google.
Just to be clear; I'm in favor of ditching the include Flash by default in Chrome, but I'm not in favor of them making exceptions for certain sites. That's just horse shit.
It does seem like a rather arbitrary decision, but they only get 1 year extension (hopefully there won't be some "last-minute compromises", to allow say the top 3 sites use flash for yet another year).
Because something something, DO NO EVIL, something something.
Seriously, Google is as bad as every other tech company, but for some reason, when they do shit like this, people tend to over look it (not sure if because a lot of folks directly or indirectly have their livelihoods connected to Google properties or what not)
I'm rather baffled as to the sinister narrative you're attempting to construct here.
(I sometimes worry I come across as a Google apologist. I suppose I am in a devil's advocate sense. It's not that I think Google is a shining beacon of decency. It's just that the quality of anti-Google thinking on HN is so low I find it difficult to let it pass. If you could all step up your game somewhat then I can stop chiming in).
I notice the same thing. As far as the Flash thing, I'd imagine it's a matter of trying to balance the desire to wean people off of using insecure plugins while minimizing complaints. Certainly not "fair" to give some a pass and not others but at the same time, if (made up number) 95% of non-sketchy Flash usage takes place on ten sites, you can weed out a huge swath of exploit problems while minimizing hassle as you wait for the top 10 to move away from Flash as well.
It's an imperfect solution to be sure. But in these situations (officially ending support for WinXP/IE6, ending support for some plugin, etc) there is no perfect one. There needs to be some motivation to get users and websites to move on from Flash but there's also consideration for breaking commonly used services.
YouTube is the only Google property on the list. In fact, there are more Amazon properties than Google. So the "point" you were trying to make didn't work out.
My point is that they are building html5 versions of many of their sites (like Youtube, probably the biggest site supporting html5). Why would they care about serving the old flash versions of those sites to their html5 browser?
I agree that they have not moved all of their properties in lockstep over to html5. But I think Youtube supporting html5 provides more information about their motivations than the number of domains they are putting on the default whitelist.
As others have mentioned, Google finance is as well as well as other finance sites. Makes you wonder how someone haven't introduced any HTML5 graphs for stats yet...
I started to watch a lot of youtube videos recently, but it doesnt seem all of them require EME. Am I wrong? Because it seems I can download them fine with a firefox addon, except a very few of them for obvious copyright reasons. I also rarely see song videos that cannot be downloaded from mp3 conversion websites. No idea if those websites are legal or not.
Anyway, on one side you have youtube advertising and adblockers, on the other side they want to force people to stay online and not use the content offline, but they can't have all of them not block ads.
I really wonder about the real cost of hosting youtube videos, and the money they make from ads on youtube. I think there is a huge loss there. Youtube RED might solve this.
I hate it when I see people talking about Flash as if it's the Web's cancer. Remember Flash pushed the web heavily forward back when we had to install all sorts of desktop applications and codecs just to watch a video or play a simple game. It is an optional plugin that you can always disable. Not having it at all would lead to a poorer UX in my opinion. Can't Chrome just have a toggle Flash on/off in the toolbar and get over it? And let Flash die naturally, when HTML5 catches up?
Edit
Also the propaganda [1] against it makes me laugh. The reason Google wants it out is probably because it can't track the ads inside Flash apps.
Hell I could rant about this for long. And some people who say Flash is insecure (it probably is) are also the ones who install ad-blockers, antiviruses and all sorts of creepy browser extensions. Well that s not safe either, but Steve Jobs said Flash is bad so it must be. However Steve Jobs knew he cannot monetize the web as well as the AppStore so the "open" web was pulled backwards.
ok, are Google planning on supporting HLS if they're ditching flash? I see this site is saying that it does (with their player) -- so does it natively or are these guys doing a transmux in Javascript similar to HLS.js and others?
Chrome supports HLS natively on Android, but not desktop. Safari and Edge support HLS natively on desktop. Firefox doesn't support HLS natively on Android or desktop.
Later this year we plan to change how Chromium hints to websites about the presence of Flash Player, by changing the default response of Navigator.plugins and Navigator.mimeTypes. If a site offers an HTML5 experience, this change will make that the primary experience. We will continue to ship Flash Player with Chrome, and if a site truly requires Flash, a prompt will appear at the top of the page when the user first visits that site, giving them the option of allowing it to run for that site (see the proposal[2] for the mock-ups). [1]
"By the end of this year Chrome will begin ignoring Flash as Google takes another step towards removing the “final Plugin” by replacing Flash with HTML5."
DRM has never been important for online video workflow and it's only useful in the way a walled garden is useful... it will require that each pair of browser-website (with EME) negotiate before allowing it, basically making it impossible for smaller browsers to become useful at all