It's incredible how many people are sympathetic towards Google and Apple's opposition of Ogg Theora within the HTML 5 codec debate. Of course, nobody expected anything else from Apple, but Google, really? Outside of YouTube, Google seems to be very supportive. YouTube may still be somewhat separated from the rest of Google. The site still looks more like it did pre-aquisition than it looks like a Google project. The Google Chrome web browser, a "pure" Google project, supports Theora, and a blog post Google made on the Ogg Theora book sprint which specifically makes mention of HTML 5 leads me to wonder if they might actually be planning on supporting it. Then again, this claim by a Googler on the WhatWG mailing list makes me worry:
So far, only one large video host, Dailymotion, supports HTML 5 with Theora, as well as The Video Bay. In addition, there is also Wikimedia Commons, The Internet Archive, and Tinyvid, but it doesn't end there. Ogg Thoera and Vorbis are also supported by big projects like OLPC, Jabber, the Mozilla Foundation, and the Wikimedia Foundation. Their support is vital in showing that these formats are a viable and good choice for multimedia, but we still need the big dogs with the big bucks.
Now, what can we do to push Youtube along with other large video hosts and advertising services to move towards full HTML5 and Theora support? If you use YouTube or another major video host that doesn't offer HTML 5 with Theora, post a video explaining your support. We need organized effective ways for people to convince these sites to publish in free formats. Please post additional ideas and feedback in the comments.
"If [youtube] were to switch to theora and maintain even a semblance of the current youtube quality it would take up most available bandwidth across the Internet." --Chris DiBonaAll right, it's time to fight back against this FUD and address every argument made against Ogg Theora for HTML 5 and YouTube. The whole point of the HTML 5 video tag is to eliminate the need for proprietary Flash plugins so are we really going to replace one proprietary piece of technology with yet another piece of proprietary technology? What would be the point of that? Have we not already seen enough of the problems posed by relying on proprietary standards? Then again, HTML 4 didn't mandate support for JPEG and PNG, so gaining support doesn't necessarily depend on becoming the official standard.
- Video quality (and high definition)
- Firstly, even if Ogg Theora's video quality was less than ideal, this is a web standard we're discussing here. It isn't about the highest quality format of the moment. The purpose of a standard is to allow everyone to take part and rely on it. A patent encumbered format does no good for this. Freedom and transparency are essential for a standard. We wouldn't have the world wide web or TCP/IP networks if competing companies were in a constant battle trying to outdo each other's protocols and hardware. Nobody is forced to use the HTML 5 standard; if Apple is so afraid of Theora's success, they can go ahead and use their format of choice or continue to rely on Flash.
- Secondly, Theroa will continue improving, and will improve faster if it gains more support. Open formats like this can evolve for a reason. Video quality will improve, just like the HTML standard itself has grown and changed. Agreeing on an open format for a standard is done to promote mass adoption which allows for widespread use unencumbered by patents and without similar restrictions. If a proprietary format is used, who will improve upon it? Only the
intellectualimaginary property "owners" will be able to, and that doesn't mean they ever will. Theora might not be the very best right now, but it will be. - Finally, although Theora is not yet at the same level as H.264, claiming that its quality is so inferior is a lie. Not only is it not far behind, but it is already better than what YouTube currently uses. What about high definition, you ask? Although it is still not ideal for HD content, being slightly behind H.264, Theora roughly equals YouTube HD video quality already. Most of the problems it had were inherited from the codec it is based off of, and those problems are disappearing thanks to support from from the Mozilla and Wikimedia Foundations with the upcoming version 1.1 of Theora. This is the overhaul containing a major rewrite bringing big improvements and lots of new features like two-pass encoding and adaptive quantization. Theora has a balance between quality, performance, size, and simplicity which makes it ideal for web video.
- Patents
- To refuse to implement a royalty free codec under the guise of "patent safety" is ridiculous. Apple wouldn't bundle Samba with OSX if they were that afraid of patents. Yes, submarine patents could possibly exist, but then again they are always a possibility, and not just for Theora, but for many formats. The algorithms Theora uses are not new. If, down the road, some patent-troll does in fact appear, whatever patent they hold would likely expire when brought to light.
- Hardware Acceleration
- Yeah, yeah, Apple is understandably against using Theora becuse of this, but this isn't much of a reason against it becoming a standard. Again, keep in mind the most important thing is not the absolute best quality and performance, although Theora is already pretty good as it is, but rather freedom and independence from proprietary technologies.
- Most of the affordable portable media players, "MP4 Players", support Vorbis without advertising it. The sole manufacturer of all the chips simply included it. It might take a while, but hardware support for Theora will come, and we could see it sooner if it became more popular. There is already an open source VHDL code base for a hardware Theora decoder in development and it's worth mentioning that decoding Theora is less CPU intensive than decoding H.264.
- Dirac, On2 and other Open Formats
- Dirac is for high definition video archiving, not streaming. Nothing is wrong with other open formats, but Theora is already the best candidate with the most support.
- Google did recently acquire On2, a company which owns more advanced codecs, and actually created the codec which Theora is based off of. There is speculation that Google will offer them as free formats, and as awesome as that would be, we can't really act on what they might do.
- Everyone already has Flash
- This is probably the dumbest argument i've heard. Of course, Flash isn't going anywhere overnight, but online video will shift to use the HTML 5 video tag. A quarter of the world's web browsers already support Ogg Theora with no plugins required. YouTube and other sites won't just drop Flash support one day; like Dailymotion, they will begin by offering HTML 5 video in addition to Flash. From there, HTML 5 gives us so many possibilities: no additional software needed, easily created player skins, dynamic content injection, and the ability to manipulate videos, just to name a few.
So far, only one large video host, Dailymotion, supports HTML 5 with Theora, as well as The Video Bay. In addition, there is also Wikimedia Commons, The Internet Archive, and Tinyvid, but it doesn't end there. Ogg Thoera and Vorbis are also supported by big projects like OLPC, Jabber, the Mozilla Foundation, and the Wikimedia Foundation. Their support is vital in showing that these formats are a viable and good choice for multimedia, but we still need the big dogs with the big bucks.
Now, what can we do to push Youtube along with other large video hosts and advertising services to move towards full HTML5 and Theora support? If you use YouTube or another major video host that doesn't offer HTML 5 with Theora, post a video explaining your support. We need organized effective ways for people to convince these sites to publish in free formats. Please post additional ideas and feedback in the comments.
34 comments:
Amen.
Until there's full screen video, flash isn't going anywhere. This just the way things will be! I'm sorry. People don't care what plays the videos as long as it works they want it to.
Besides personally I would rather they wait until HTML5 becomes and unofficial standard. (By which point hopefully there is fullscreen video added to the spec, and also hopefully by then the browsers can sort out their cpu usage on slow ercomputers when playing HTML5 videos)
The Google claim that OGG would increase traffic in a unbearable way is easy to fix: just make youtube more cache friendly!
With OGG it is easy to enable proxy caches like Squid to cache the video stream and serve it in the local network or autonomous system, in the same way images and static files are handled by those caches. And certainly, all ISPs would enjoy a lot being able to do so.
So choosing an open and free standard like Theora (OGG Video) would end up in SAVING bandwith instead of consuming more.
@Anonymous #2
For Linux users both those problems can be solved at the same time. Full screen video should call out to a dedicated video player.
The video pipeline in browsers (and Flash) is slowed by the need to be able to place arbitrary text images etc on top of the video. Fullscreen doesn't need that.
The Ogg Theora ecosystem is still at the stage where it just needs geeks (of all stripes) to use it.
So use it. Bug others in Linux or FLOSS communities to support it wherever possible. Figure out where it's broke, or could be tweaked or streamlined or improved. Then improve it, or report the issue to someone who can. Repeat.
Note also that if Google releases the codec it just bought, then it still needs an audio codec and a container to work with it. Ogg and Vorbis are pretty good candidates, so many of the same process and tools will continue to work.
There are two main reasons youtube and flash video so popular:
1. Near instant start, no waiting to watch.
2. The pre installation of flash player on many computers.
Normal people don't care about video quality or anything else, and they don't want to install stuff.
I think you miss the point. For youtube, bandwidth is a *major* cost. They spend a truckload of cash each month, and work closely with Adobe to work on formats which deliver a high quality/bandwidth ratio.
Asking them to adopt Theora is asking them to spend literally millions of dollars of month on what is already a money-losing service to them. If you want Youtube to move to Theora, prove that it would deliver the same quality with less bandwidth, or better quality with the same bandwidth. Then the numbers would speak for themselves and Google would be more likely to experiment with Theora.
Ryan, this argument is addressed in Danny's very first point above. Do read the link he provided; encoding to theora would cost extra storage and cpu just like supporting mp4 for the iphone does. It has no bandwidth cost at all.
What's "incredible" is the fact that you take Google's _shipping support for Ogg Theora_ in the platform it controls (Google Chrome)as "opposition". Google is highly supportive of Ogg Theora in general. It is a mistake to take one person's comments about one product (which he does not even work on) as some sort of evidence of company-wide opposition.
-- Peter Kasting, a Chromium developer
Stuff like this is just silly. "It doesn't matter if it looks like *** because it's FREE." We're talking about digital records of peoples' everyday lives. If it means a licensed encoder/decoder (especially one that can already be had for FREE [costwise] on other platforms), then people are going to prefer quality over some vague ideology that they don't feel applies to them.
There are lots of reasons that Youtube would use H.264. The first and foremost being that they already have most of their content encoded in it, and have gone to great labor to do so. H.264 is supported in hardware implementations such as the iPhone, Tivo, etc. that they are contractually bound to support. Not to mention that they can use the SAME H.264 encoded file in an HTML5 video element, and pipe it through Flash for everybody else to reach virtually everyone.
The whole point of H.264 is that it is a *single* format to handle everything from HD down to web sized. And it's HERE TODAY and already supported everywhere from the professional video industry on down to Adobe Flash. It's supported in some 98% of the world's browsers, which is a far cry from Theora's 24% in a single browser (Opera and Google don't even support it in official builds yet, do they?). There's no appealing incentive other than Theora being free for anyone to switch. I'm an open source advocate and fan, but BS like this just makes me sick.
OGG Theora is extremely high quality. I'm not sure about HD because I don't have a powerful enough computer to play those sizes, but of what content I have watched that is encoded in Theora, I say death to Flash and H.264. Oh and to Kenneth Pardue above, Google supports Theora as does Opera.
All you said makes sense if freedom is a priority. Sadly, it is not for the parties you mentioned.
I'd love to see Google serve Youtube videos in Theora. The cost of converting a billion videos to Theora is probably preventing Google from doing so. Google already has everything in h.264 for iPhones and recent versions of Flash.
Also, Google will need to keep the Flash player available for quite some time to support legacy browsers. Using h.264 means they can serve the same video file to Flash users and HTML5 users.
I think h.264 is the better codec and has more momentum. The _only_ problem with it is the possibility of lawsuits. I think it would be easier to manoeuver around those problems and there are many simpler measures that can be had.
Id say google are going to use On2's VP8 as soon as they actually acquire the company. Theora's compression needs a little bit of work but its getting there. Using MP4s for youtube would be a step backwards IMO. Also when google's buying of On2 is finalized it will remove the question marks over theora's head too.
This week's Dilbert explains the situation sufficiently:
http://www.dilbert.com/strips/comic/2009-09-02/
"Apple wouldn't bundle Samba with OSX if they were that afraid of patents."
Actually that's not quite true. The Samba team's agreement with Microsoft protects all who develop for, use and distribute Samba, including Apple.
http://samba.org/samba/PFIF/index.html
@Chris NotSoSmart:
Since when is Apple shipping Samba? And since when do patents exist? And when did the PFIF happen?
You do the math.
About quality and patents: does anybody complain about HTML only supporting crappy old jpeg? The newer JPEG2000 allows much, much better image compression, yet nobody seems to be interested in using it for the web.
Theora vs. H264 looks a little bit like JPEG vs. JPEG2000. If browser don't need JPEG2000, then they also don't need H264!
My question was whether or not Google and Opera currently support Theora as a company-wide ideological decision, specifically I was asking if it's in any officially shipping product (not a beta or developer snapshot). To my knowledge, they aren't. The point there being is that people are proclaiming that 24% of the world's browser (singular) support Theora out of the box, whereas virtually 100% supports H.264 through the video tag for browsers that support it, and through Flash for the rest. No need to have two files or depend on some clunky Java workaround. 24% isn't a pro, it's a con. End users don't see the difference between native and plugin-based video, other than that the native video doesn't go fullscreen.
The Theora vs. H.264 situation is a little bit like what Anonymous said, but reversed. JPEG vs. JPEG2000 -- correct that if the browser doesn't need JPEG2000 then so be it, but in this case it seems that H.264 is JPEG and Theora is JPEG2000. H.264 is already ingrained in the professional video industry as well as the web industry.
Take it or leave it, the reality is that Mozilla and others are simply engaging in political grandstanding with this whole Theora mess, and introducing an unwanted variable in actually getting everybody away from proprietary video into something that is a standard. Maybe not a *free* standard, but a standard that is vendor neutral, and widely implemented in 99% of the world's desktop operating systems. Heck, even Windows 7 is shipping H.264 support built in.
It's also questionable as to whether or not Theora can even be improved much more without violating patents that are held by other firms. It would be nice if all this emotional investment spent demonizing H.264 in favor of a less documented, lower quality codec it would be put towards lobbying the holders of the members of the patent pool to provide a royalty-free software license on H.264, and leave the hardware decoders (for Blu-ray, decoding chips, etc.) in place.
That Dilbert comic also isn't very applicable. Like it or not, Apple is not the sole proprietor of H.264. There are literally dozens of academic institutions and video companies that have put a lot of time and money into developing H.264 that are a part of the patent pool. And those guys, preferring to eat, would understandably like to get some return on their work.
Excuse me if I sound ignorant on the issue, but to an outsider, the whole Theora/HTML5 issue just seems like a solution looking for a problem. The reality is 99% of computers can stream video today after installing a piece of proprietary (but FREE) software that in all likelihood is already installed on the computer. Given that, IN REALITY, what problem does Theora solve? Not requiring this <1 minute Flash/Silverlight installation?
Let's just say hypothetically that Flash and Silverlight were embedded in the major browsers, would this change the situation? Or would people still be wanting a video tag with a common codec? I'm just trying to understand the problem...
Everything I have read about Theora says that it is TERRIBLE in terms of quality/MB when compared to H.264. Not to mention that H.264 _already_ has all of the advanced features that this article claims that Theora *might have* one day. A video clip of comparable quality ends up being 30%+ larger in size when encoded with Theora. I realize there are more complexities when comparing the outputs, but I don't think anyone would argue the general statement.
It's irrational to make YouTube double their bandwidth bill overnight for absolutely no gain (probably a loss). H.264 is superior in every way technologically. The single ONLY reason to use Theora over H.264 that I can tell is this "open standard" stuff which I'm simply not convinced is worth the trouble (at this point in time). As a software developer and as a user, I'm not willing to sacrifice the quality of video/audio just to make something "open". Part of the problem is that when dealing with "open standards", the argument quickly becomes philosophical rather than practical. Philosophically, yes, a universal codec would be a great thing. But practically, it doesn't make much sense given the current state of the technology and the fact that H.264 already has tons of momentum because it's already being used for almost all HD content and a lot of SD content as well.
Realize what is being proposed. They want everyone to use this other codec that basically no one uses currently. Not only that, they are asking people to replace their content with inferior technology for a THEORAtical (pun intended) gain. I say theoretical because there is no actual proof that anyone will benefit from Theora being used.
YouTube's arguments about bandwidth are completely reasonable. Even today I experience buffering and inconsistencies (albeit very rare) while on YouTube. These issues would only be magnified. So rather than smoother playback in a proprietary player, we get laggy playback in a ubiquitous player? I realize this is an exaggeration, but the point is still valid. While broadband is VERY common nowadays, cellular/3G connections are getting more and more common and bandwidth is not unlimited. It's also important to recognize the scale on which streaming video sites operate. Even a percentage point difference in bandwidth requirements and video size can make a huge difference on the overall performance of the system.
Asking geeks to use this isn't a good idea either. Geeks are all about technology. There is no way I'm willing to sacrifice the quality of the codec that I already have. In fact it's the complete opposite. My suggestion is for Theora to first catch up to H.264 technologically, and then come back and we can talk about replacing H.264 as "THE" codec.
@Kenneth Pardue
You just don't know what you are talking about.
Clunky java workaround?
Not even JS is needed:
http://camendesign.com/code/video_for_everybody
Tom: I beg to differ. The video for everybody workaround is even less practical than Java, requiring that video be encoded in, guess what, H.264 ALSO in order to also truly work for everybody. Maintaining two sets of encodes for a large number of videos is completely impractical, especially when a single H.264 file will work fine. I mean, really, you're telling me that I don't know what I'm talking about because there's an even LESS efficient solution out there that I didn't mention? :-P
Zac's comments are spot on. Youtube may be big, but maintaining a copy of all of their videos in H.264 to satisfy contracts and the remaining 76% of the world that sits on planet reality AND Theora for the fringe groups, simply isn't practical. And I suspect that his sentiments reflect most average web users: why would I want to sacrifice quality for some vague concept? Theora really is a solution (and a crappy one at that) in search of a problem.
One thing that I will say is that having a common codec actually IS practical. It reduces headaches, confusion, and makes it easier to implement across the board. Fortunately, Apple has shipped with H.264 for years, and Microsoft will beginning with Windows 7. Not to mention that H.264 can be had in Adobe's free Flash player or Microsoft's Silverlight, the former of which is also available on Linux. Despite the concerns over freedom, we were very close to settling on H.264 before Mozilla started all this Theora is the Savior of the Internet crap. Before that, nobody could give a flip. Mozilla is the prime reason that HTML5 video will not take off as quickly as it otherwise might have.
I really can't see an upside to Theora besides the philosophical one.
Myths are not debunked with opinion, sorry.
Hi,
This is really nice post.I get some good facts about the Ogg theora and HTML 5.The Youtube matter will be solved soon by the Google.
hdmi adapter
It is interesting to know about Ogg theora and HTML 5.The Youtube matter will be solved soon by the Google.
Good interesting post, thanks for sharing.
Although...I don't think you do give quite enough weight to the arguments against Theora. It really will be a problem for Google to transcode and serve all their video content in Theora, and Apple probably do have some genuine concerns over the legality of implementing Theora in their software.
However, Kenneth and Zac, you really don't seem to have any understanding of the point of Open Source. It's not just philosophical ideology. Mozilla Firefox, not to mention Chromium - the project behind Google Chrome - are open source project so that people who want to can download the code and play around with them in any way that they want, and redistribute. Without worrying about being prosecuted. This may not mean anything to the end user, but it is essential for development in the browser, and Mozilla is not ever going to ship with H.264 support as long as it itself is not open source - because then they could no longer allow people to download and tinker with Firefox's code. This is why Google Chrome supports H.264 but Chromium *does not*.
And to respond to Zac's comment about why do we need a <video> element instead of just continuing to use Flash in the browser, the answer is very similar. Flash doesn't integrate at all with the browser, or the rest of the web page. This means that web developers can't play around with video within their web applications. I understand this doesn't currently matter to the end user, but developers will be able to do all sorts of clever things with <video> in the browser which was never possible with Flash.
In both cases, end users might not care at this stage, but should the supporters of HTML5 and Theora have their way, I guarantee, the end users will be glad of the results.
Robin.
http://people.xiph.org/~maikmerten/youtube/
<--- Why Theora is better.... (note that is from the dev branch of Theora aka libtheora 1.1)
it this changes ONLY apples to the encoding, so no need to re encode videos...
This is for the fullscreen comments (which actually made sense). Mozilla Firefox 3.6 already supports fullscreen theora :D
So they apparently support it.
But just kind of.
See this hilarious.. conversion failure: http://www.youtube.com/watch?v=Zse0Y0zsvbY
In case they fix it before you read this: http://yfrog.com/jq201002082104241089x1177p
I'd like to get a smartphone that is full of open-sourced/free hardware and software, which means they are free from patents. This means I get to have a cheaper product which does not compromise too heavily on quality. I would pay only for the cost of the product, plus additional labor costs to build the product.
For example, I'd get a smartphone with Symbian, Theora+Vorbis, Wi-Fi, Garmin, etc. The only reason I avoided buying myself a HTC was because it uses Windows Embedded (which I guess isn't free). I'd rather trade that extra cash which I need to pay for Windows Embedded for some other open-sourced feature such as Theora.
@Kenneth Pardue
I'd buy Mozilla's "Theora is the Savior of the Internet" argument, simply because it makes things cheaper and as good (if not better) for the consumer.
If H.264 were to become a standard, each smartphone vendor, smartphone user, PDA vendor, PDA user, web developer, web user, basically almost everybody in this modern world, would have to dig into their pockets to make themselves poorer, and at the same time, making 1 already-powerful entity, the MPEG-LA, yet richer and yet more powerful.
I'm a startup doing surveillance cameras. I don't think my business would take off if I had to use H.264, and for me, Theora was the obvious choice.
Well, I don't think the Internet would have taken off anyway, if TCP/IP, FTP, HTTP, SSL, etc., were all patent-encumbered.
Just look at what PNG has done to GIF (which was previously patent-encumbered, until the patent expired a few years ago). PNG has sort of become the de facto standard in web graphics, has better quality than GIF, better compression than GIF, and it's free.
Post a Comment