This is awesome news. No patent encumbrance means Firefox will support it. Terrific performance means the rest of us will actually use it. Google backing the technology means Youtube, with their unfathomably large proportion of the world's video content, will support it.
Not really. Google might have paid for a copy of Flash or Flex to build the player, but other than that they don't pay any licensing fees to Adobe since they don't use Flash media server.
Well, it seems like Google can easily add the codec support to WebKit, in which case it would be hard for Apple not to support it -- they'd have to explicitly disable that code. That seems unlikely given that Apple has really been pushing HTML5 forward, and it's no work from them for a cool feature.
Obviously getting Microsoft to add codec support is another issue altogether, but it seems plausible that it could be done. Since Google would own all of the IP, they can sign some patent contract with Microsoft promising that yes, this really is open, and no, they won't ever sue Microsoft over implementing the standard. If that were the case it would make the VP8 codec the most promising among the alternatives for Microsoft, since it would be free and its IP status would be certain and guaranteed. It seems like the question hinges on whether or not Microsoft plans to add <video> support at all, and since they've made vague claims supporting HTML5 and these new fangled web standards, it seems plausible that some version of Internet Explorer could support the codec in the future (even if it takes them a while).
Er, Google just controls the Chrome and Android ports of WebKit, not the Mac, Windows, or iPhone ports. They all have their own media frameworks which are separate; the Mac and Windows Safari ports use QuickTime, and thus support whatever codecs QuickTime supports (so, you can install a Theora plugin, and presumably VP8 once it's released), while Chrome uses a modified version of ffmpeg, and I'm not sure what the iPhone or Android ports use.
Getting Adobe Flash to support it would be a much cooler business hack, and would probably force Apple and Microsoft's hand in to the bargain. (And they have been cosying up recently)
To be honest it's getting away from flash which makes this exciting for me. Wrapping VP8 in Flash would be the equivalent of putting granite wheels on a Ferrari.
No, it would be like if you wanted to put really nice wheels (VP8) on your Ferrari (Chrome) but those wheels won't take you far if the roads haven't been upgraded (widespread usage of VP8 on websites). So you upgrade the tires of all of the crappy cars (IE6,7,8 and probably 9) for free and without them having to do anything (Flash auto-update to VP8) and then people start upgrading the roads (websites using VP8). Most people have still got crappy cars, so they might not notice any improvement but it makes a big difference to everyone else trying to make better cars and roads (the open web).
Giving up on the car metaphor, it's Flash support for H.264 that makes it viable for 85% of the market. Adobe can win it for VP8 if they want to, but I think Google can do it with out them too, it'll just take longer.
Well, except that Firefox already has a non-patent-encumbered format they've put a lot of effort into. Apple/Nokia refused to implement the previous non-patent-encumbered format, why would they change their mind for this one?
I'm not sure Google ever actually stated on-the-record why they refused to add Theora support to YouTube, but 'cost to re-encode all that video' has been widely cited, and I don't see why re-encoding with VP8 would be particularly cheaper than re-encoding with Theora.
Theora had several drawbacks. Its performance is simply not as good as h264. It's a previous-generation codec based on VP3, whereas VP8 is cutting edge.
Flash never supported Theora, which invalidates its use for the vast majority of web users today. Even with HTML5 video, content protection worries is going to keep a lot of content wrapped up in Flash players for a while. On the other hand, Flash has a good chance to support VP8--Adobe already holds a license option on VP7 and has supported the codec line in their players.
The lack of players, content hosts, and embedded decoders all prevented Theora from ever taking off. I think there's a better chance that Google's direct backing of a codec (both as a content host, mobile platform, and browser, give VP8 a better shot.
I have also read that Theora's "patent-free" status is questionable; to my knowledge nobody has reviewed the codebase to ascertain what patents it may infringe on. On2, as a commercial enterprise licensing its codecs to others, has had to navigate the patent minefield already. I trust its status more, especially with Google behind it.
As for what Firefox will do, well, that remains to be seen. I just think it makes sense for VP8 to succeed where Theora failed.
Theora is not as good as h264... but neither is VP8 (granted, VP8 is closer to h264, but it's still inferior).
I didn't know that Adobe had an option on VP7, but I'd be impressed if Adobe added VP8 support to Flash - I mean, the whole point of this exercise is to undermine Flash, and I'd be surprised if Adobe joined in except in the face of overwhelming customer demand.
Yes - if Google officially supports VP8 in Chrome, Android and YouTube, that'll be a very strong argument (far stronger than Theora has had) for other content providers and user-agents to support it... but Apple/Nokia/MS and YouTube already support h264, so adding support for another, less efficient codec seems silly.
Theora's "patent-free" status comes entirely from On2 saying "we give everybody the right to use all our IP involved in VP3". Presumably On2 had the same due-diligence issues when it was licensing VP3 to others; I don't see how On2 could give VP8 a stronger "patent-free" declaration than Theora already has, unless (as other people have suggested) Google has bankrolled an exhaustive search for applicable patents.
I don't really mind whether VP8 or Theora winds up being the dominant format on the web, I'm just skeptical that VP8 is all that different when it comes to the pain-points that have prevented people from adopting Theora.
In the last Theora/H.264 thread on HN, I recall somebody saying that VP8 was better than H.264... when VP8 was new, and H.264 encoders were still largely experimental. Nowdays, things like x264 and commercial codecs shipped with QuickTime are better than VP8.
Actually that author/developer usually specifically calls out Apple's Quicktime as a very poor H.264 implementation, so if he says VP8 will be nearly as good as the second best H.264 implementation then he's saying that it will be much better than Apple's. That probably applies to Adobe's encoder and in fact the vast majority of commercial H.264 encoders.
Considering he's an x264 developer that article is very positive about the quality of the VP codecs. The stuff he's down on e.g. it being proprietary, licence fees, patents, big name corporate approval, mobile support, buggy encoder are all stuff people expect Google to fix quite easily (well apart from patents).
(My earlier reply was to the original version of your comment.)
It will benefit Google in two ways:
First, it insulates them from encoder licensing fees going forward, and from decoder license fees if MPEG-LA decides to back off their free-for-web-streaming arrangement. It also means they're free to bundle the codec with Chrome.
Second, if you believe the publishing hype, it could save them up to 50% bandwidth. I doubt it's that effective, but at Youtube's scale, processing is cheap and bandwidth is expensive; every 1% helps. If they can reduce the number of storage nodes, bytes transferred, and by extension, open conns/sec, that translates to very real savings.
They won't recode from h264 and drop the originals immediately. There are too many hardware h264 devices out there. My guess (given how their transition from h263 went) is that youtube will add it as another encoding target and use it for their web formats as soon as flash supports it. Mobile will remain h264 in 3gp, but that's only one out of six formats.
Why is everybody talking about recoding? Can't they just keep the current videos as they are and encode new things in a different codec? As far as I know all the old H263 videos on Youtube have never been recoded. And why would they? Recoding H263 to H264 would just degrade quality.
I understand what you're saying, and I'd love to be educated if you know more about the topic, but aren't they both perceptual encoders? Ie, wouldn't the parts of each frame they're identifying as unimportant be quite similar?
(I'm quite certain I'm wrong on this point, but I'm interested in learning why)
I understand the sentence, but what I'm trying to understand is why the information removed is (significantly) different - since they're both perceptual codecs, there should be a large overlap in what is removed from the output.
Eg, take a frame which has a solid chunk of blue. Codec 1 removes some of data to make things more consistent and then compresses it. Codec 2 looks at the output of codec 1, sees the consistently large blue area, doesn't see any info worth removing, and compresses it.
I don't know all the maths behind it, but try saving a JPG, close it, open it, save it again, repeat a few times. You'll see quality degrade every time you close it. My guess is that the lossy compressors try their best to preserve the previous compression artifacts, but in doing so they introduce more compression artifacts.
You pay per encoder and per decoder, both of which would adversely affect Google with their encoding farms and popular browser if the charge wasn't capped at around $5 million (I think it's gone up to just under 6 million this year) to stop the big players doing something silly like getting together and developing their own royalty-free codec.
Unfortunately for that cunning plan, Google clearly sees more benefit from a codec that's compatible with the open web and open source than they do from licence discounts.
Kudos to Google for being smart enough to realize that an open web is a win for them in the long run and for having the guts to compete on the strength of their products and not on platform/api lock-in. Obviously it's still an uphill battle to get this adopted across the board but they're setting the right example here.
I can't find any official word that Firefox is supporting it. This article simply says it expects Firefox's will because of Firefox's reason for not supporting h.264 was that it wasn't open license.
Given how much caterwauling there's been about Firefox not supporting H.264 and the problems with Theora, I'd be very surprised if Firefox doesn't support VP8 soon after it's open-sourced. Too many people want it too badly for it not to happen.
At launch, On2 went so far as to claim that it could provide “50 percent bandwidth savings compared to leading H.264 implementations.”
Considering the amount of bandwidth Google spends on streaming YouTube content, this move is far from altruistic. The more viewers adopt a bandwidth-superior format, the less terabytes of data to stream, the less the cost of running YouTube. They have absolutely no benefit from any kind of "platform/api lockin" to H.264 or any other video standard, only that there are less bits to stream.
I don't see how it's an "uphill battle" to convince anyone they should use a free standard instead of paying million of dollars in licensing fees per year. I'd expect Microsoft to ship this instead of H.264 in the next major IE, and, if this comes with a good and open hardware interpreter, Apple to follow soon after.
Considering the amount of bandwidth Google spends on streaming YouTube content, this move is far from altruistic.
Who cares? Companies are allowed to make the world a better place without having to self-flagellate themselves. Google could easily have just paid the H.264 licensing fees and told Mozilla to toss off, or even built VP8 support into Chrome and used YouTube to force-feed it to the world. They chose to cooperate, and should be praised for that.
I don't see how it's an "uphill battle" to convince anyone they should use a free standard instead of paying million of dollars in licensing fees
Several companies, most notably Apple, have already spent what must be millions of dollars on licensing fees and hardware. Regardless of which is a better long-term choice, it can be very difficult to reverse that much momentum. Not to mention the hundreds of millions of non-consumer multimedia equipment which is designed to use H.264.
Additionally, VP8 does not have much application support yet. It'll take a while for codecs to be written, debugged, and distributed to users. There may be significant lag time before VP8 displaces Theora among F/OSS users, or H.264 on OS X.
They chose the path that would lead to the fastest adoption of the technology which lowers their bandwidth costs. Keeping the technology Chrome-only is against their financial interests.
Apple and MS pay yearly licensing fees for H.264. As soon as they adopt VP8, they can stop.
Considering the amount of bandwidth Google spends on streaming YouTube content, this move is far from altruistic. The more viewers adopt a bandwidth-superior format, the less terabytes of data to stream, the less the cost of running YouTube.
While I'm not arguing with the "less bandwidth is good" argument, I think it's a very common misconception that the amount of bandwidth YouTube uses is a significant financial strain on Google.
For example, on the YouTube blog Google have said:
The truth is that all our infrastructure is built from scratch, which means models that use standard industry pricing are too high when it comes to bandwidth and similar costs. We are at a point where growth is definitely good for our bottom line, not bad.http://ytbizblog.blogspot.com/2009/07/youtube-myth-busting.h...
Wired wrote a piece where they interviewed a network analyst who thinks Google's bandwidth bill is close to zero:
"I think Google’s transit costs are close to zero," said Craig Labovitz, the chief scientist for Arbor Networks and a longtime internet researcher. Arbor Networks, which sells network monitoring equipment used by about 70 percent of the net’s ISPs, likely knows more about the net’s ebbs and flows than anyone outside of the National Security Agency.http://www.wired.com/epicenter/2009/10/youtube-bandwidth/
No one will remove H.264 - it's a great codec with wide support. It's more likely that we'll be in a world with two good video codecs, not just one. VP8 will hopefully become the HTML5 video standard, and H.264 will remain in Flash (which will still be useful for a long time), most browsers, many mobile devices, and on the desktop.
Incidentally, this has been the state of audio for a long time - mp3 and aac are both good codecs, and both are really widely supported. (Vorbis is reasonably good too, though not as widely supported. Maybe that will change with HTML5.)
I'm glad that their interest of reaching a wider user base to advertise too is also in sync with creating an open platform like the web. They can keep the Ad platform secret, nobody will care as long as they keep doing things like this.
This is basically a done deal, fwiw. Other folks who I've talked to in a nda-ish sort of manner have been expecting to not support OGG long term, and instead support VP8 as the "open standard." It's an open secret in industry circles, and this I think confirms it. Can't wait till next week.
Are there any chips out there that do hardware decoding of VP8? That's a critical path item to get Apple on board - they've learned from Nintendo; battery life is too precious to sacrifice for a mere codec.
VP8 support is currently mostly nonexistent. It's currently just a minor proprietary codec that almost no one cares about or uses. Of course, Google certainly has the money and resources to fix this relatively quickly if they really want to.
On2 has press claiming that VP8 is "40% less complex" than H.264, which presumably means that it can be decoded more efficiently than H.264 by some benchmark or another. There's probably some sort of marketing spin to this, but if it's true that VP8 is as efficient or more so than H.264, then Google opening up the IP to write VP8 decodes would probably mean that a lot of hardware vendors would start writing support for it.
Yes there are chips for VP8 decode and they're very popular. They're called ARM.
VP8 was specifically designed by On2 (in collaboration with ARM) to be decoded easily on low powered ARM chips. They take advantage of various ARM features such as NEON acceleration and are intended to work well even on lowly ARM9, and the codec has built in scalability that lets low-powered devices show the same video file that everyone else gets with less quality and power if needed.
The hardware argument was basically an excuse for Apple to not have to do something they didn't want to when talking about Theora. It's been shown to be irrelevant, or even a benefit to Theora when sending video to people who don't have state of the art laptops. To apply it to VP8 is just ridiculous.
A chip like an OMAP2 or OMAP3 has a built-in DSP used to assist and accelerate video decoding. They've already added H.264 decoding to the IVA for some devices (hi, Dell) so adding new bridge code for VP8 is certainly possible.
Apple's a big backer of H.264, and they're invested in the codec. Hopefully they'll support VP8 in Safari, but I wouldn't look for the iPhone to change to VP8.
Is it possible for existing H.264 hardware decoder chips to be used to (even partially) help decode VP8? Or will we see a new generation of chips that does H.264 and VP8?
Potentially. Theora activists keep repeating that "there's no such thing as a hardware decoder; they're all programmable DSPs", but in addition to programmable hardware you also need the desire and resources to actually implement VP8 decoding firmware for all these different chips.
(Odd, in all the HTML5 video debates on HN, I hadn't heard that argument from Theora supporters before.) Most of the hardware decoders in smartphones are programmable DSPs; e.g. TI's OMAP3 family. But there still are plenty of chips out there that use fixed-function h.264 decoders, e.g. TI's DaVinci family. The fixed-function ones are smaller, cheaper, and probably lower power.
TI's DaVinci family includes programmable DSP and ARM based solutions and you can buy various 3rd party software implementations of codecs such as H.264 to run on them.
Ok I am going to say it: I think this finally secures a big big future for Chrome.
It will now be the only browser to support all three of the main codecs - and Google will control one of the two open source / royalty free ones (indeed it's arguable that status may be more secure that Ogg Theora because, simply, Google are likely to aggressively defend any submarine attacks).
This is pure speculation but maybe Google uses VP8 in order to negiotiate a better licensing fee once h.264 isn't free anymore. MPEG-LA would think twice about loosing Youtube as a customer. This would go against Google's "Don't be evil" but getting the world to accept yet another video codec will be an uphill struggle, even for Google.
Surely making the codec open source isn't in itself that interesting, considering that x264 and FFmpeg form a high quality, open source codec for H.264? It's not the code, but the spec that matters in spreading adoption of VP8. (Of course, publishing the code would create a de facto spec).
More interesting still is the patent issue. VP8 is still effectively irrelevant unless Google announces they've performed a patent search and found VP8 to have no known patents covering its methods (or at least any that might incur licensing fees). After all, it was the patent issue with H.264 that bothered Mozilla.
x264 is a codec implementation, not a codec standard. The x264 source code is free, but what it produces isn't (at least in the US). Rumors are VP8 will be free on both counts.
As for the patent issue, Google has no doubt done this during its ~9 months of due diligence while buying On2. Not saying that issues won't come up, but Google apparently felt comfortable with the situation.
I am confused as to why more people aren't asking this. If On2 were licensing any patents (for example, from one or more of the MPEGLA partners) in order to make VP8 possible, it's not important that Google has open sourced it unless they also somehow provide a free license to all the patents involved.
I don't know enough about the industry to be able to predict whether On2 could have a technology in VP8 that they, and only they, have patents on. In that case, the ball is in Google's court. Otherwise, whoever owns the patents that On2 was licensing (or infringing?) still as a right to get royalties, right? Experts?
As the article points out, the big wildcard is Microsoft. I don't see what logical reason Microsoft has for skipping VP8 other than the anti-Google animosity, but that has led to illogical decisions on their part before.
By now I tend to just assume that Internet Explorer is going to always do the bare minimum needed to maintain its dominant position, while holding back support for whatever people want. I hate it when I develop a cool web app and it works perfectly on everything except IE, where it's broken in new and excitingly different ways on each of the major versions.
I haven't tried this in IE for a while, but can't we do something like push down an ActiveX control or plugin to render this video with IE pretty trivially? It would simply become part of the magic incantation web developers would use to embed video.
Flash did not obtain its dominance by coming installed with IE.
Unless Microsoft actively moves to block it, I don't know that Microsoft's explicit support is actually that interesting.
/Flash did not obtain its dominance by coming installed with IE./
That's exactly how Flash won its dominance. It was installed by the PC manufacturers (HP, Apple, Dell, &c) into IE5 and IE6 until it had overwhelming market presence.
Adding video support later on cemented the position, but it was strippers and steaks (for HP, Dell, Apple, &c execs) that won Flash its position.
If they can make sure H.264 become the standard, then that's a big problem for firefox, which is good for Microsoft, so it wouldn't be totally illogical for them to skip VP8.
There is a sample video of On2 on this page:
http://video.golem.de/audio-video/2857/on2-vergleich-h.264-u...
Text basically says that those are crops of FullHD movies.. keep in mind that this movie is from On2, so its validity is at least questionable ;)
This may not solve the problem, as companies like Apple, Microsoft, Adobe, Nokia, etc need to get on board as well. They are likely to continue the same arguments they raised against Theora; most importantly that VP8 is still vulnerable to patent extortion despite being independently developed. It's probably too much to ask for Google to provide indemnification from patent lawsuits, but hopefully they've at least done some work to help assuage these concerns.
Adobe is already shipping VP6 as a part of Flash Player, so VP8 should not be a problem. Maybe that is what Google will get for integrating Flash into Chrome, that made news few weeks ago.
MS, Apple and Nokia will be harder nut to crack, because all three of them are patent owners in H.264 portfolio. Nokia may cave when Youtube switches, maybe even MS, but Apple users will have life slightly more difficult for a while.
Given that theora is based on On2's VP3, could there be enough similarity in the codecs that the theora DSP code is useful for quickly getting VP8 going?
Is anyone besides me worried that Steve Jobs is going to say "they tried to kill the iphone, so we will never adopt their evil standards!"? Is the animosity between google and apple enough to stifle development?
yesbabyyes is not suggesting that Google will force Apple in to adopting VP8, he is just pointing out the fact that if Apple doesn't support VP8 their users won't be able to watch YouTube videos. Google won't have to do anything, Apple consumers will pressure Apple to adopt this standard in order to be able to play YouTube videos.
I'm not worried about that at all. Rejecting web-standards that others like Firefox and Chrome are implementing is surely a path to destruction and I don't think SJ is that crazy.
Nitpick, but video codec is not a web standard. There was a big discussion whether HTML5 should require support for some codec, and in the end it does not.
Otherwise Apple's WebKit (which is the rendering engine for Chrome) is doing very well in terms of web-standards.
"There was a big discussion whether HTML5 should require support for some codec, and in the end it does not."
This could cause them to reconsider that decision. In fact I'd go so far as to say this should cause them to reconsider their decision. Perhaps they'll come to the same one in the end, but this is a radical landscape change and it is appropriate for the standards committee to react to that and not just passively say "Oh well, the standard is written in stone, so sad."
If there's any consistent theme to the history of web standards, it's that all the successful standards start as someone's proprietary extension that eventually becomes a de facto standard before being codified by the standards committee. Standards that start in the committee rarely gain traction.
Ah, true, I was a little hasty. Still I don't see how Apple could realistically refuse to support this in Safari, since presumably support for codecs is directly built into WebKit and either Apple or Google would have to fork it if they wanted to go in separate directions. That's not totally out of the question though I suppose...
Apple and Nokia refused to support Theora, and the same argument applies to VP8 since it's also from the VP family. Google would probably need to provide indemnification to change Apple's mind.
Safari and Chrome both use WebKit but they use completely different code to implement <video>; Safari uses QuickTime and Chrome uses ffmpeg.
Apple and Nokia refused to support Theora, because they have vested interest in H.264 - they both are on the licensors side. Why would they support free codec, when they get cut from every license fee paid for the encumbered one?
The argument used by both of them is just convenient excuse. What do you expect of them? To tell you what is in the above paragraph?
Jobs is a businessman, it would make sense to him to try and avoid or reduce licensing costs. Google turned from a close partner to being a competitor, not an enemy.
I'm not seing any confirmation from Google for now or any other sources. This is probably going to happen anyway, but for now I won't take it as a sure thing.
It's kind of funny how having false publicity for it now has probably amped up expectations enough that Google would look bad if they didn't follow through.
Sweet.