Submitted by This warrants a new post as far as I’m concerned, mostly because the original post is getting buried in updates and will soon drop below the fold. Microsoft has just announced it will support VP8 in HTML5 video in Internet Explorer 9, but only if the user has the DirectShow filter installed. Update: Yes, the updates keep on coming. Zencoder has added support for VP8. Update II: Zencoder’s side project, video.js, offers a player that can fallback between h.264, OGG and VP8 on most browsers. Support for Android browsers is underway too. Update III: The H264 supporters’ hardware argument for mobile is sounding moot too, since ARM explains on its blog that mobile devices with Cortex-A8 and Snapdragon processors “will be able to take advantage of WebM” through those chips’ NEON SIMD engine.
This is a major shift from what Microsoft said before about HTML5 video in Internet Explorer 9. When the company detailed its HTML5 video support, the company was very adamant in that it would go all-in on H264. It would be the only video codec supported by Internet Explorer 9, no matter which codecs the user had installed. Most likely pressured by the prospect of a VP8/WebM-only YouTube, Microsoft has now given in.
“In its HTML5 support, IE9 will support playback of H.264 video as well as VP8 video when the user has installed a VP8 codec on Windows,” writes Dean Hachamovitch, General Manager, Internet Explorer. This puts support for VP8/WebM in Internet Explorer 9 on the same level as support for H264, with the only caveat being that Windows ships with H264 out of the box, whereas support for VP8/WebM has to be installed separately by the user.
This leaves only Apple’s Safari out of the loop. While Google announced plugins for Gstreamer, and has already made the DirectShow filters available, there’s no peep yet about QuickTime plugins (Safari’s video falls back on whatever QuickTime can decode). It will likely be up to a third party or Apple itself to implement support for VP8 in QuickTime.
Times are getting interesting.
As much as I don’t like MS I have to congratulate them for this change of direction.
I think the strong demand for really open formats is very important and organizations like Mozilla and Wikimedia and others play a really important role in this. (And thanks again, Google)
And that makes Apple the only bastion of patented closed crap. But Steve Jobs is so full of crap that does not really wonder me that much. They will take a long time to change. They hate openness (although the open web enables the Macs return) and it will be their downfall (eventually).
But Apple is supporting many and many open standards and open source projects !
Look at LLVM, OpenCL, OpenGL …
CUPS man.. they own the X of printer support.
If you are trying to say that they own the source code for CUPS you are correct (through acquisition around 2007?). This does not make Apple open!!
No, but it means they hold ownership of a pretty big open source component along with the others mentioned in the previous post. They are far from an open company and surely not a poster candidate for the FOSS marketing campaign but they retain open licenses on some pretty serious chunks of code. The printing framework is akin to the kernel, a GUI framework like X, .. it’s not a small dependency or obscure user app with several booming alternatives.
Supporting Open source projects & Open and royalty free standards is different.
Steve Jobs wanted to support open standards for web. They are not at all interested in the HTML5 part. All they are interested is the H.264 part since Apple owns one patent in MPEGLA’s H.264 patent pool.
OTOH open source is a different thing.
Apple I would argue was the first to really push into HTML5 (on the browser front)…. to this day, they have the furthest implementation of it … nothing to do with H.264
I can’t speak for the current HTML5 implementation, but Apple have been marginally behind Opera for much of the ACID3 testing (Opera actually scored 100% before webkit/Safari did).
But I do agree that Apple have also been front runners.
Apple has interest in H.264 more than HTML5.
They are pushing HTML5 basically for H.264. If they really want to be so open, then why don’t they open up the iPhone/iPad app distribution model and allow side-loading of apps?
HTML5 has nothing to do with H.264. The W3C originally specified Theora as the video codec for HTML5, but Apple threw a hissy fit (and Google said that Theora wasn’t good enough for YouTube, which was true at the time), and so the W3C had to take Theora out of the HTML5 spec because of lack of consensus.
HTML5 now simply fails to mention any codec. It does not recommend H.264.
Perhaps now that Google have made WebM another open, royalty-free video codec option, and Google are happy with that performance, then W3C can now fix HTML5 and make it specify WebM instead, an Apple consensus be dammed (you can’t please everybody).
Edited 2010-05-20 13:01 UTC
Apple may use and support open standards and open source projects but it doesn’t respect its users. Apple feels its users should be lucky they got the opportunity to spend their money on Apple products.
as far as i know, quicktime handles the video tag in html5 and will use whatever codec the tag specifies if the codec is already installed on the OS. I believe safari currently can handle theora if the user install the codec and the codec already exist, can somebody confirm this? Other than on their devices, apple doesnt seem to actively prevent other codecs and i dont see a reason why they wont go VP8 in the future. Apple has control issues but i dont see how they will be affected by this.
Microsoft said IE9 wont even bother to check what codec the user has, its h.264 only and now they seem to have changed their minds and they are saying they will support VP8 when the tag specifies it and it is installed. One can assume they will also support theora if installed too since the architecture will already be in place
No. They have said very clearly that HTML5 video in IE9 will be restricted to H264 and now VP8. That’s it.
Thom is right. IE9 will only support those 2 (but notice i did say IE9…)
Which is very sensible if you ask me. Video codecs are one of the least secure classes of software.
@boulabiar
Apple is odd that it has some decent open source projects like Grand Central Dispatch and Webkit (forked off KHTML), and Darwin (not so much these days)so they seem like a company that is very open.
but then at the same time have some of the most locked down products that exist on the planet making them look very controlling and walled off.
Mixed feelings. =/
Edited 2010-05-19 19:31 UTC
Having a few opensource projects doesnt make Apple ‘open’.
Also, Webkit, the most-mentioned apple foss project, is foss for the following reasons:
1) Its a fork of khtml which is LGPL. They could not have changed the license.
2) whole webkit’s success was because of wider adoption of it by third parties. if it was closed, it wasnt going to have any success.
and that doesnt make apple open.
Lets be honest, KHTML wasn’t going anywhere and was probably about to be abandoned since Firefox worked a lot better on *nix.
Apple probably looked at all their options, including licensing Tasman, Trident or Presto, using Gecko, or writing their own engine (closed or open). Things could have gone differently. No one forced Apple to pick KHTML to fork from.
That is absolute FUD. KHTML was the most standards compliant engine at the time.
Which didn’t help it any when it couldn’t render half of the sites people wanted to get to without borkign the pages. Not that it was its fault, the environment was different then, but for anyone but Apple to see a gem in KHTML at the time would have been rare. I used to avoid it like the plague, every site I went to looked like crap. Text was off, things didn’t align correctly etc. Maybe its standards compliance is why Apple decided to go for it, but KHTML of old and Webkit are totally different animals now. Apple made it into an engine that anyone can use while before it was heavily tied into QT.
@apoclypse
That’s where I was going with that thought.
Yeah, as you said they probably have looked at all their options, but yet decided to use khtml. because it was great. And it costs them. They dont have total control over webkit because they chose to fork khtml and not write it from scratch.
Citation please. From where you heard this? KHTML was one of the finest rendering engine. I understand it wasn’t perfect, but still that powered Konquerer and throwing away and rendering engine is not a sane idea.
I would like to know from where you heard this?
From a technology standpoint nothing was wrong with it.
The problem was no one was using it, and because of the environment of the web it didn’t render most web pages correctly. If Apple didn’t fork it can you honestly tell me you think KHTML would be around today?
WTH?
Why you think apple’s fork helped khtml?
I remember the discussions on KDE planet and forums. Webkit actually reduced people working on KHTML. All the Qt folks moved to Webkit and some KDE folks too. KHTML would have been better than what it is now if Webkit was not there. Would it be better than Webkit… definitely NO.
Well, your citation is still missing
That’s insane. Webkit’s success is because Webkit is GREAT. That’s WHY third parties picked it.
It doesn’t make Apple open but your Logic is flawed. Apple could have created their own web engine if they chose to and/or licensed a commercial one. They chose KHTML because they saw some merit in the technology and chose to go with it despite it being LGPL. There are a lot of instances where Apple could have gone with the more closed solution and instead opted for the more open solution. If the LGPL was an issue for Apple they wouldn’t have used it in the first place. 3rd party adoption is precisely because a huge company like Apple is actively developing the engine and throwing their resources behind it. Webkit isn’t even the smallest or most memory efficient rendering engine out there, but it has support from a company that actively develops it and its opensource.
Oh please. You are judging his logic but you just put your own spin on the same facts.
1) Writing an entire rendering engine would have been exponentially more expensive to develop and take years longer.
2) Apple did not pick KHTML because it was open source.
3) By your own admission Apple saw merit in KHTML _at minimum_ as a good base for Webkit.
While KHTML might not be as advanced as it is today without Webkit, Webkit would be a largely irrelevant curiousity without KHTML. It would be nothing more than another proprietary engine running in Apple’s shroud of secrecy. Writing a rendering engine from scratch takes years of focused development – even for Apple. To devalue KHTML so implicitly is just emotional blindness.
Apple turned a solid open source project into a fine rendering engine. Both parties deserve equal credit, especially since one party didn’t earn a dime off it.
I think your attempts in presenting a rational line of thinking is, … well…wasted on the original poster. The same poster defends H.264 and everything that Apple does, to the point of posting things like “despite being licensed under LPGL” (wtf???) or ignoring all other contributors to webkit.
As for the news … couldn’t have arrived at a better time, also is it just me or is this a bit of wave (with long reaching ripple effects
That’s true. And if I remember well, KHTML devs had to threaten Apple before they released the source ^^
That’s one of the reasons why Apple loves BSD licensing : GPL variants are pretty annoying for the proprietary world…
This is not a change of direction. They lost the game.
Fact 1) MS needs to keep its browser in competetion.
Fact 2) Youtube is ‘the’ internet tv. Users want it no matter what
Fact 3) Youtube is owned by Google.
Fact 4) Google also owns IE’s competeitor, Chrome.
Fact 5) if IE becomes incompatible with youtube, it will die.
Fact 6) See Fact 1.
What im saying is that if IE decides not to support VP8, Google can change youtube’s videos to VP8 and kill IE.
Microsoft doesnt take such risk.
Please stop ‘Microsoft changed their mind’ stuff. They had to.
Edited 2010-05-19 19:40 UTC
They still changed there minds and direction no matter t motive
Agreed. Microsoft no longer have the dominance to dictate standards to the content providers – particularly given that Google is one of those providers. As far as HTML5 is concerned, I expect MS to play nice – they need IE9 to keep them relevant in the browser space, which won’t happen if they go against a united opposition.
WHY does MS need to keep their browser in competition? They do not make any money directly off it, it and it would probably be cheaper to just pay off mozilla or some other browser to use Bing and other MS web services by default rather than pay for the development (and everything else) surrounding IE.
It is about control. Microsoft have fought a decades-long battle to keep SVG from being widely adopted by refusing to support SVG in IE. Similar story (but not quite) for PNG, where Microsoft took many years before they offered any more than broken, partial suport of PNG. Same for DOM2. Same for a correct and performant implementation of ECMAscript.
Microsoft don’t want open, interoperable, anyone-can-implement standard formats and protocols to become widely used. If they were to be widely used, people wouldn’t need to run a Microsoft platform any longer in order to use closed, unique-to-Microsoft formats and protocols instead.
Edited 2010-05-21 03:20 UTC
Flash 10.1 will support WebM, so IE9 can just wait it out for the update, if needed. But when IE9 released, Flash 10.1 will probably be on most machines already.
Again, IE’s future will depend on some other company: Adobe.
Thanks ! It’s easier to understand this move, now.
However, isn’t it a little fast ? VP8 isn’t even on some YouTube videos yet !
Technically Microsoft will not be supporting anything. You will still have to install the DirectShow filter.
And you have to read the text carefully, they said that “IE9 will support”* not “Microsoft will support”.
* – As much as IE9 will support any other codec as long as DirectShow filter is there.
For the billionth time, THAT IS NOT TRUE. Microsoft has made it VERY clear that ONLY H264 and now VP8 will be supported by IE9 – other codecs are NOT supported, whether you have the codec installed or not.
This isn’t rocket science, people.
H.264 is built in, VP8 and Ogg/Theora are optional via DirectShow plugins. Nothing has changed.
All 3 posts at http://blogs.msdn.com/ie/default.aspx have nothing about exclusivity of HTML5 video codec. Just that IE9 will have H.264 built in.
You are right, this is not rocket science, it’s simple click and read.
Doh!
No general directShow, so no Theora or any other codec. ONLY H.264 and now ONLY VP8.
Is the concept of the word ONLY so hard to understand?
This is wrong. H264 is NOT built into Internet Explorer 9 at all. IE9 simply makes use of the H264 Media Foundation codec (not DirectShow), and IE9 has been RESTRICTED to H264 SPECIFICALLY. Microsoft has made that very clear in their posts.
And, more specifically (emphasis mine):
This is all very clear.
Something did change; namely, that Microsoft has added VP8 as the second permitted HTML5 video codec in IE9, alongside H264.
There is no technical reason why IE9 could not support other DirectShow/Media Foundation codecs as well (e.g. how Safari simply uses whatever QuickTime supports), but Microsoft has opted to artificially limited HTML5 video codec support in IE9 to H264 and now VP8.
This comes straight from Microsoft itself, and can be verified all over the web. Why continue to say otherwise?
Edited 2010-05-20 00:11 UTC
I am glad MS went through with this, it literally almost came down to a coin toss on MS’s part.
Edited 2010-05-19 19:11 UTC
hmm they only said about vp8. what about entire webm, especially mkv container and vorbis audio?
At the end of the day, YouTube is like the porn film industry: It gets to decide what format lives and what format dies.
If Google decides to go full-VP8 with YouTube, every browser maker will have to add support for it or fall back to Flash. This puts Apple in a very tight position: It’s clear that they don’t want Adobe in the game but will they implement VP8 into QuickTime?
Ah, I need some popcorn.
It is a popular belief that the Porn Industry killed Betamax, but it absolutely isn’t a fact. From what I’ve read VHS was already more or less ‘established’, when the Porn Industry chose to back it.
Although many concidered Betamax being superior to VHS, there where a few defining characteristics that made Betamax the less favorable choice for consumers:
– limited recording capacity of 1 hour versus VHS’s 3 hours
– higher manufacturing costs and thus higher consumer price
– VHS was easier to use
A few links on the topic:
http://www.mediacollege.com/video/format/compare/betamax-vhs.html
http://en.wikipedia.org/wiki/Betamax ( doesn’t even mention the Porn Industry )
Edited 2010-05-19 20:53 UTC
Point remains that without porn driving innovation, we’d still be using 9600 baud modems, viewing gif images, still ordering things over the phone and relying on VHS for distributing video.
I’m joking, but actually, I’m not. The porn industry was one of the early drivers for things like streaming video, online transactions and high-bandwidth applications. Much like VHS, the fact that these innovations also had legitimate use and were therefore entirely justifiable for one to have without labeling one’s self as a deviant, doesn’t change the fact that porn was a huge driver for adoption.
In fact, I would be surprised to not see the porn industry adopt webm for online video. They could save the licensing fees, and frankly I suspect they may hold more sway than YouTube on viewing requirements. They would virtually force adoption by the browser holdouts, because a browser incompatible with porn is a browser with a very limited future. Even if nobody wants to admit using it for that.
Be a good way for the porn industry to get back at Apple for disparaging them.
How is this any different from the way things are today?
youtube.com is (currently) a Flash site. Anyone with the Flash plugin, regardless of OS, can access it. Behind the scenes, the videos are in a variety of formats.
If they went VP8-only tomorrow, and Adobe releases a Flash plugin with VP8 support tomorrow, nobody would notice any difference.
The only way Google/youtube can dictate the video format is if they go VP8-only … and forbid Adobe from adding VP8 support to Flash. Which they are not doing.
Thus, business as usual for everyone … including MacOS X users.
Not only are Google not forbidding Adobe to add VP8 support to Flas, Google have actually partnered with Adobe for this WebM project.
http://blogs.adobe.com/flashplatform/2010/05/adobe_support_for_vp8….
http://mashable.com/2010/05/19/vp8-webm-support/
http://www.engadget.com/2010/05/19/google-launches-open-webm-web-vi…
The only significant party in web video who has NOT offered any support for WebM is Apple. Even Microsoft have backed down from their earlier position where they said that IE9 would support ONLY h.264, now they say IE9 will support WebM also.
This is obviously a huge development. Without IE9 on the WebM bandwagon, video support on different browsers would be a huge mess. Safari and IE9 would support only H.264, Opera and Firefox Theora and VP8 and Chrome everything. That would mean that anyone who wanted to publish a video would need to encode at least twice.
Now (when IE9 is released and VP8 gets in the release builds) every browser but Safari can play VP8 video. Even for Safari, VP8 is just a plug-in away and I imagine we will have one very soon. That means that there isn’t really a reason to encode to anything but VP8.
Of course the situation is less than optimal since IE9 and Safari users would need to install codecs, but if WebM catches on, I’m pretty sure that we will see support out of the box in Windows 8 and 10.7.
WebM will not be there… It will come faster with Flash 10.1. But IE9 will only support VP8/WebM vie DirectShow filter/plugin.
I now realize that the statement was a bit ambiguous. I didn’t mean the release builds of IE9. I meant the release builds of Chrome, Firefox and Opera as opposed to nightly/lab builds.
I see no mention on that blog of native support for WebM. Heck, I don’t even see WebM mentioned on there!? A user who has Chrome installed can play WebM without installing anything. I don’t think that’s the case with IE9. VP8 != WebM.
http://www.webmproject.org/about/faq/#vp8_video_codec
As an aside, I bet my coworker $20 that Apple won’t introduce support for WebM in Safari.
Edited 2010-05-19 19:36 UTC
And video.js uses VfE!
Happy day for me.
Thrilling! All the browsers I’d actually use support it now – I’m happy. Hopefully Safari will get on board… even though I don’t like it personally I know folks who do.
It would be strange if Safari wouldn’t join. If Youtube (it’s just one site but an important one) becomes VP8 and HTML5 by default Safari would have to fall back to Flash. Which would be a little hypocritical with there stance against Flash.
… would be the first hypocritical thing they did, but still …
http://www.webmproject.org/about/faq/
“In the coming weeks we will release QuickTime and DirectShow plug-ins that will enable many third-party encoding applications to produce WebM files.”
Now my question would be if VP8 or WebM does any patent infringement, who is going to protect whom ?
If Google or WebM takes care of everything, then this is really very good news.
Does WebM (VP8) support hardware acceleration ?
If Yes, this is awwsome! otherwise, battery on all mobile devices are gone.
Is VP8 royalty free license irrevocable ?
If Yes another awwsome! (RedBull)
Can Xiph.org kill Vp3/Theora now ?
If they are doing , it is good only. To avoid confusion for users which one to choose and less maintenence cost.
Finally, As I always said only google with YouTube has power to change the world in all the comment posting that OsNews did on HTML5 Video. It came true. YouTube has become part of Internet life, that no one can kill it. Virtually they own the HTML5 Video tag
At the end, seems like all are happy
Thanks a lot Google
yes.
yes
No. work continues on theora.
So does this mean that video hosting sites will need to store 3x the amount of data to ensure videos will be played?
With the Flash update, this could be changed to use only HTML5/Webm and fall back to Flash+WebM if not supported. H.264 would have to be included too for Apple / IE.
Thanks google (again)… I may not be a fan of ms and especially ie but this was a good move on their part and it puts MORE PRESSURE on the remaining holdouts. One small step for goole one giant leap for the open web! YESSSSSSSSSSSSSSSSSSS!
bashing of Microsoft over this issue, it was Google that has monopoly power to push whatever codec they want, not Microsoft. And this proves that. That’s why I never understood the ire at Microsoft. And the notion that they were making huge money from the H264 patent pool was flat out false.
I’ll add that Microsoft’s IE team (by far, the WORST team they have, for years now), still doesn’t get it. They should just toss the video playing responsibilities to the OS and be done with it. If I want to host an HTML5 page using DivX, VC-1, MPEG2, Sorenson QT, DV, or even MJPEG (whatever it’s called) for that matter, why shouldn’t I be able to do that? Not that I would, but there’s nothing preventing IE9 from being totally agnostic about what codecs to support. It’s not like it takes more code to be agnostic; in fact it takes less code.
I hear this a lot, but there is one very good reason *not* to farm all video tags off to the OS’ built-in decoders:
Security.
By allowing any page to host any video format supported by the underlying OS, it could allow any website to host malicious audio/video streams that exploit yet-unknown bugs in common codecs that area already available on many machines.
By reducing the supported video formats, you reduce the exposure, and can concentrate on securing only the codecs that are supported.
This doesn’t actually stop someone from forcing a user to download a malicious audio/video file anyway, but the “barrier to entry” is higher when the user doesn’t get the video embedded directly in the browser as they expect.
I could be way off base, but the fact that each of these browser developers is baking the codec *into* the browser suggests they may care very much about this issue rather than installing/using codecs already available on the host platform (For example, Chrome could also automatically install VP8 for DirectShow, Quicktime, or GStreamer, and just use it from there rather than including it directly in the codebase).
Considering the fact that IE is now sandboxed, I don’t really see this as much of a threat, though.
Codecs are native code running directly on the CPU… and unless Microsoft has found a way to prevent “dangerous” code from running within a codec, there is always a potential for exploitable code execution.
True but it’s not really an argument since the problem is easily prevented by only allowing certain codecs to be used and handled with the OS API’s.
I’m sure the browser makers have their reasons but I’m also sure none of them are about security.
NEON SIMD is like x86 SSE and it is not a hardware video decoder. All codecs use SIMD when available, that does not make them hardware accelerated. And some ARM chips have a video h264 encoder which would be outright impossible to implement with reasonable quality/compression even with SIMD code.