Now that Internet Explorer 9 has been let out its cage, we all know a great deal more about Microsoft’s position towards the video codec situation with the HTML5 video tag. Microsoft has chosen for H264, a codec it already includes in Windows by default anyway. This means that apart from Firefox and Opera, every other major browser will support H264. Some are seeing this as a reason for Mozilla to give in to their ideals and include support for H264 as well – I say: Mozilla, stick to your ideals. The last people you should be listening to in matters like this are web developers.
The problem is, however, that most people don’t care. So if Firefox can’t play H.264, and thus can’t watch Youtube and insert video site of choice here, the users aren’t going to shun Youtube but will shun Firefox. Simple as that. Look, this isn’t bloody complicated. On platforms that have a multimedia framework, use it! Theora can be played internally, while other videos will be spawned off to the framework. Mozilla keeps their ideals, and end users keep Firefox instead of switching to Chrome or IE. Unless I’m mistaken, Opera has done this at least with GStreamer on Linux. Mozilla’s current stance is going to lose them a huge market share.
Actually, I think EVERY SINGLE OS that Opera has a current version for has a media framework.
And, the OSes that Mozilla maintains current versions for have media frameworks.
Windows: DirectShow
Mac OS X: QuickTime
Linux: GStreamer
FreeBSD: GStreamer
Solaris: GStreamer
Of course, I’ve heard the claim that Firefox actually does support GStreamer… but only uses it for Theora.
Indeed, but I’d only heard mention of Opera supporting Gstreamer. Whether Directshow or Quicktime support is in Opera I’m not sure, heck I’m not even sure the Gstreamer support is fully implemented yet.
The GStreamer support is fully implemented in Opera 10.50 (which is released for Windows, in beta for OS X and in alpha for Unix), but they don’t use DirectShow or QuickTime. In Linux/BSD they use the existing GStreamer installation, so Opera can play anything that GStreamer can handle (including h264 if you have the codec installed).
In Windows and OS X they install a minimal installation of GStreamer that can only handle Vorbis, Theora and PCM. As for adding other codecs on Win/Mac you can’t do it as you would in Linux/BSD, since Opera’s version of GStreamer misses a lot of things, including the plug-in system.
But since GStreamer is licensed under the LGPL, Opera has made available the code for their version, so it could be possible to directly link the plug-ins with the library used by Opera.
Actually Opera is first developer to port gstremer to non-unix environmnet. Becuase they had not so much time they ported only most important thing, is using DirectX to display image and audio (but actuall encoding is done in gstremer). If they had more time, or support from gstremer developer (which are esentially all Linux guys) they could easly support other codecs.
They are probably doing this already, or done it for Windows, but are not shiping with H.264 enabled both for licensing issues, and for political reasons. On Linux it (h264) works in Opera, and esentially codec stuff are much easier to por than core framework. codec is mostly some CPU algorithms, which will just easly recompile to ther target. Core stuff is harder as it interacts with whole system (threading, windowing system, etc.).
This is not a technical problem to support all this stuff now. It is political.
Actually Opera isn’t the first to port GStreamer to Windows. The first release of the GStreamer WinBuilds project was on late 2008. As for using DirectX, this is just how GStreamer works. It doesn’t interface itself with the hardware, but it interfaces with another API (OSS or ALSA in Linux) that interfaces with the actual hardware. The API that speaks to the hardware in Windows is DirectX, so that’s what they used. And, as you said, not porting enough of GStreamer to support h264 was a political decision and had nothing to do with Opera’s timeframe. A developer said that supporting every codec under the sun wasn’t their aim and that’s why their version of GStreamer only supports their core codecs (Vorbis, Theora, PCM). Supporting more that these on systems that already have GStreamer is just a side effect.
Indeed, and Songbird uses GStreamer on all platforms as well.
Interesting, and also very bizarre as there is a Directshow plugin, not sure about a QuickTime one.
I was baffled too initially, since the have done the same thing for their UI (custom UI framework with win32, .Net, carbon, cocoa, qt, gtk+ etc. backends (some of these new, some of these deprecated)), but if you look at their goals it makes sense. Why go through the trouble of designing a framework and all the backends, if you only need a very small subset of what these backends can do?
By the way, QuickTime can handle H.264 natively and xiph.org provides the XiphQT QuickTime component for Vorbis and Theora.
YouTube isn’t going to cost itself money, be totally at the whim of MPEG LA, and throw away all Firefox and Opera users.
Dailymotion and Wikipedia have a role here too.
http://videoonwikipedia.org/index.html
You will need a Theora player in your browser to see all of this content.
What possible reason could Microsoft and Apple have for not providing a capability to render Theora/HTML5? It is not as if including a Theora player is going to cost them anything. If they don’t render HTML5/Theora, IE and Safari will, once again, be seen as the browsers that lack features and short-change their users.
They both hold some of the patents in the patent soup that makes up h.264 – and thus collect payments for licenses of h.264. It’s all business (though some keep saying “politics” – it’s business, they are not the same).
Safari does have the capability, I just went to that site and watched some of the videos. What it doesn’t do is support it out the box. Install the Xiph codec and it works. Perian might also provide Theroa support as it wraps up ffmpeg, I’m not sure.
That’s just part of the problem.
I’m assuming that H264 chips need drivers. I’m assuming those drivers will not be available as Free/open source. I’m assuming that even if they were, they’d be patent-encumbered too.
It goes so much deeper than you think. We’re just going to exchange one proprietary video delivery method with another. It makes no sense to rail on Flash, but promote H264. It is so incredibly inconsistent.
GPUs these days are programmable.
http://en.wikipedia.org/wiki/Gpgpu
http://en.wikipedia.org/wiki/Video_Shader
One can easily implement hardware video acceleration for Theora using features of GPUs such as these.
h264 does NOT have a monopoly on hardware video acceleration for any device large enough to include a GPU.
He’s not referring to GPGPU decoding of H.264, he’s referring to dedicated H.264 decode hardware. That is, silicon that implements an H.264 decoder.
And, that’s actually one thing that hurts Theora – to target mobile devices, there’s not enough CPU (or GPU) to decode Theora in software, and the only decode silicon is for H.264 and maybe H.263.
The CPU grunt required to decode Theora is much less than h264. As Theora gets better and better at compression, this can go either way … there is less data to process, but it may require more processing.
In any event, many devices that struggle with h264 can handle Theora. Any kind of 2D hardware acceleration helps, even as simple as Xv.
http://en.wikipedia.org/wiki/X_video_extension
Virtually all graphics hardware supports this level of assistance to the CPU, and that is often enough for Theora.
Edited 2010-03-19 15:17 UTC
Except:
1) Many devices include dedicated hardware to decode h.264 – none include dedicated hardware to decode Theora. THIS IS EXTREMELY IMPORTANT AND CANNOT BE DISMISSED.
2) The OGG container is not suitable for streaming video.
I don’t really care which codec wins any more. I thought Theora was a poorer choice, but it seems that the issue is really the OGG container format is crippling Theora as a streamed entity. If this was addressed, I’ve no doubt that Theora is a viable codec.
Edited 2010-03-19 16:48 UTC
Ah, you read the ranting blog post by Mans Rullgard [x264 developer] that got posted to Slashdot as a ‘story’ and formed some pretty strong opinions based on that. Apparently Ogg is so bad that it couldn’t possbily be working well as the sole video delivery mechanism for Firefox and Opera and Dailymotion and Wikipedia… everyone must be imagining it.
Let’s just say Mans’s assessment wasn’t exactly accurate. It had just enough big words and made up numbers to seem credible.
That would be too expensive. It’s usually just a DSP slightly tailored to “typical video decoding tasks” that then gets some code upload to actually know the codec.
There are several reasons why it’s done that way:
1. h264 is friggin’ complex. You don’t shove _all_ that in silicon.
2. You won’t be able to support updates/changes (mpeg standards have the nasty habit of being extended by additional profiles all the time)
3. You can’t deliver the chip without a patent license.
3. is especially important for the highly integrated chips (think SoCs) where you might not want to pay the mpeg licenses for every chip. And don’t want to build two editions of the chip (with and without video accelerator) either. Just don’t provide the firmware, and it’s not an mpeg device.
Now, given that there is firmware to upload, given the demand, there can be updates even for new formats. Theora and MPEG share some (ancient) traits of codec design, so at least some parts of Theora should fit onto the DSP. There you have it: market demand -> Theora decoding in hardware.
While it’s unlikely that this would happen for old devices (Too much of a throw-away society), having to develop a dedicated new chip would deter hardware vendors from adding codecs to newer revisions. Adding some code won’t, esp. if it’s – for all we know – patent free value-add.
http://opensource.samsungmobile.com/download/OpenSource/SPH-M900_Op…
contains a linux kernel and a couple of additional modules. Among them, a module called “mfc” (for “multi function codec”). There’s a large file containing a large amount of hex numbers. That’s the codec. Someone with an assembler for that chip function would be able to update it (and it’s segmented access, so adding more code is no problem)
So, create market demand, get a free codec. By caving in to the H264 mafia, you won’t.
The situation was/is different for MP3 players. MP3 is of a sufficiently low complexity to be handled completely in hardware (except for some basic wrapping/unwrapping, probably).
Actually, Ogg is best at streaming, that is what it was originally designed for. It isn’t quite as good at progressive downloading, which is what YouTube et al do (although it is “good enough” for that, and indexes are on the way).
Where is that silicon you speak off? I haven’t seen hardware H.264 that didn’t require firmware upload first.
(and I’ll annoy all of you with this until this rumors is either better supported or gone. See further below in this thread)
It is even possible to accelerate other codecs with embedded hardware, at least in some cases: http://www.schleef.org/blog/2009/11/11/theora-on-ti-c64x-dsp-and-om…
Of course, Flash is either VP6 (which I don’t think anyone uses any more) or H.264.
So, H.264 is the status quo.
You’ll get no argument from me about the long run where H.264 is concerned. I agree fully, it’s the wrong way to go if we want an open web. Unfortunately, the big players determine the long run, and end-users are extremely fickle and could care less about the long term ideal. Unfortunately, if Mozilla doesn’t compromise and the situation continues as is, it is Mozilla who will lose out. It’s sad, it’s unfortunate, but it’s true, and most users don’t have any kind of loyalty to Firefox. They’ll jump to Chrome or Safari, or even back to IE, in a split second if they can’t watch the content they wish to watch anymore.
still the key driver behind internet evolution.
End users will care, a lot, if MPEG-LA comes knocking on their door asking for license payments, which is exactly what MPEG-LA has said it will do, if necessary.
Also, end users will care a lot if they have to start paying for browsers (which are licensing h.264), else be forced to use crappy Internet Explorer.
Basically, h.264 is already really, really expensive to license. And MPEG-LA has stated quite clearly that when the current license term is up they are going to jack up the price and license more aggressively.
The big mega-corps won’t care, because they have the deep pockets. And the price is worth it to them because they can make the web less open, more proprietary, and that helps them.
But it screws over end users and developers alike.
People who can’t see this are blind. Just look at what Microsoft and Apple do – they try to achieve lock-in anyway possible. They want the web to be beholden to them. Having an expensive h.264 license works right into their strategy.
Wake up people.
Of course they *will care*. But they don’t care *now*, and most average users’ grasp of this situation is exactly zero. Face it, most people don’t care about anything that might happen tomorrow, they only care about today. Sure, they’ll be furious if MPEGLA comes asking them for money… but that hasn’t happened yet, so to most it’s not reality. In the mean time, all they care about is watching Youtube, and will jump to any solution that allows it if Firefox no longer will. I care about this, as do most others on this forum. We are, however, a minority of all computer users out there. If Mozilla doesn’t want to become irrelevant, they don’t really have a choice as the landscape stands at this moment in time. No one is asking them to license H.264, just link to the media framework if the os has one, and then whether the video plays or not is entirely dependent on the user. You can’t argue freedom on one hand while forcing a choice on the other. It’s a rather interesting paradox.
No, if sites don’t support Firefox for streamed video, people blame the site at least as much as the browser, until they have an opinion either way. When I see popular news sites using Windows Media Player for streamed video, obviously I think “that’s retarded, I’m never going to visit your site again.” OK, so I use Linux. But even before I was using Linux, if a site used badly-performing media plugins (Windows Media Player always performed badly in my experience), I blame the site as much as the software. I’m sure I’m not exceptional. There’s no reason people should blame the browser more than the site. Of course, if Youtube should actually not work in firefox, there’d be a problem.
there is no pragmatism in forcing mozilla to go for h264. that’s just a peer-pressure lockin attempt.
i suspect mpeg-la royalty hunt will effectively prevent h264 from being the major player in html5 video contest.
ok, now i’m getting a bit confused, so let’s try to make it clear :
how will it affect people who make their own h264 videos and publish them online? does it mean you have to have a
– os with h264 licence
– video editor with h264 licence (to encode)
– video website with h264 licence (to host)
– browser with h264 licence (to watch)
to fully comply with the h264 licensing rules? or maybe will also need a personal h264 license too?
what about sites that would transcode the video into another format? will they have to pay as well?
You will have to have a license everywhere (although there is a short moratorium on this requirement for web client, because MPEG LA realise they haven’t cornered that market just yet).
It will be just about the only thing on the planet where the costs of its initial production have been recouped years ago, yet every must pay and pay and pay again for using it.
No wonder the MPEG LA members are pushing h264 so hard, and disseminating FUD like crazy against Theora.
Here’s the thing.
Right now, H.264 is being used anyway, by Flash.
So, here’s the options:
Keep using H.264 Flash and support everyone with one method – no effort required for this
Use H.264 <video> for Chrome, Safari, and IE9, and H.264 Flash for Firefox and Opera – very little effort required here
Use H.264 <video> for Chrome, Safari, and IE9, and if you use Firefox or Opera, tough luck – this will kill Firefox and Opera if someone like YouTube does this, and again, very little efford required here
Use H.264 <video> for Chrome, Safari, and IE9, and Theora <video> for Firefox and Opera – this will (approximately) double the storage requirements for video hosts, driving costs up significantly
Use Theora <video> for Chrome, Firefox, and Opera, and drop support for Safari and IE9 – not actually an option, because they still need to keep H.264 videos around for mobile devices that can’t grok Theora
We know the options. Everybody does. That’s not what this is about.
This is about what option is the best IN THE LONG RUN. We KNOW the cop-out method that is good for TODAY, but what about TOMORROW?
What I’m saying is that it may not matter what’s best for tomorrow, because the people who matter will go with the cop-out decision, and leave those that decided “correctly” behind.
The correct answer may be to get a team of patent lawyers with a computer science background, and create compatible implementations of H.264 that don’t violate a single patent.
Replying to myself, but there is, of course, another answer that I almost forgot about.
VP8.
Google now owns it, and has a history of open-sourcing some stuff. Now, what would Google want with a video codec? They do kinda run a popular video site (currently using H.264,) they own a popular mobile platform (and could dictate VP8 support on that platform,) and they have a video chat service (no idea what it’s using now.)
So, if Google opens VP8, problem solved. Hardware support will come from Google pressure on silicon manufacturers, and it may be more efficient than H.264.
You are forgetting the most likely option: flash with H.264 for everything which doesn’t support H.264 in their video-tag.
No, I didn’t forget about it. Scroll up.
One can also use Cortado Java applet for playing Theora video on other browsers. This is what Wikipedia plans to do. Of course it still doesn’t cover mobile devices…
NO, NO, NO.
Use <video> for ALL browsers, even this which doesn’t support given codec or video tag at all! Then inside of video tag insert FLASH player for ALL browsers!
It is simpler, and it is actually much more robust, in case of browser changes, some less known browsers, updates, changes in the supported codecs, and is just RIGHT THING TO DO.
Thank you.
Well written Mr Thom Holwerda.
Sad to see that many have been “converted” to the rank of “h264 fanboism” and cannot see further than the edge of their nose.
In fact, I’m convinced nearly all the pro-h264 posts this forum will see are from people who did not even read this article entirely, and just argue for the sake of being a good fanboi.
It has nothing to do with fanboism, but economics. If using an ‘open standard’ drives up your hosting costs, and only reaches 1/4 of the internet audience, what’s the point? You’ll be out of business before H.264 starts charging for internet streaming. Theora doesn’t offer any benefit over H.264 in the next 5yrs, except making the owner ‘feel good’. As was pointed out in an earlier comment, you can reach ~99% of your audience using H.264, as it is supported by Flash, AND smartphones.
H.264 has been tested in the wild for more years than Theora, and provides a better user experience. Have you ever tried seeking in a Theora web video? Try Dailymotion’s or even Wikipedia’s demo. Did you know that the container format doesn’t include this information (duration length)? I’m sure you already read the article which highlights the various shortcomings of Ogg.
For the next 5 or so years, H.264 is simply the best format, hands down.
Theora has been around longer if you count the original VP3. It’s not exactly new.
Yup, all the time. It works pretty well.
It certainly does include the duration
The article was more fantasy and puppy kicking than reality. It had just enough facts sprinkled in to make it look credible if you didn’t fact-check the whole thing.
https://developer.mozilla.org/en/Configuring_servers_for_Ogg_media
Well, leaving aside that using ogginfo on any ogg contained media *always* shows you the duration of each stream… if people hate ogg so much, we have another option. It’s called Matroska. Personally I like it much better than ogg anyway, but either way it’s irrelevant to the codec argument. A container format can be substituted, but a lot of people here are saying “ogg” as if that means Theora. It doesn’t. Of course, it’s probably in some people’s best interest to fudge the issue. Theora in Matroska would work just as well as Theora in Ogg, and I at least wouldn’t object to the Matroska container being adopted as the standard container for Vorbis and Theora anyway.
Interesting that’s from Moz, but still not correct. All data is timestamped; duration is last timestamp minus first timestamp (streams can begin at an arbitrary offset).
Theora is a variable bit-rate format, therefore comparing two timestamps in a stream will not give you an estimate of its entire duration.
(for the record, yes I am a web developer)
I don’t think it’s quite as simple as blaming web developers, it’s usually their job to interpret requirements (not define them) and get them working in the best way possible using the current browsers out there.
If my employer decides it wants videos on it’s web site, and h264 gets support in 70% of their audience, and Theora gets 30%, I would have to go with h264. Either that or I would get told by the business to go with h264. In a commercial environment there would be no other way really.
It’s really up to the big players to define the standard that everyone else follows. The biggest problem we face is that youtube has not gone with Theora. If sites like youtube and wikipedia went with Theora then it would win the day because MS and others would have to make sure their browser supported youtube as it’s users would insist on it.
Wikipedia has already gone with Theora…
http://www.downloadsquad.com/2010/03/19/open-video-not-flash-is-com…
You can’t compare an operation which is funded by donations (Wikipedia) with one that operates to make a profit, such as YouTube.
This is an interesting game. There are some factors which will declare the winner:
1) Google/Youtube: Most of the video on web is being watched on youtube. If it switches to HTML5/h264, Firefox and Opera will have to get the license.
On the other hand, if it switches to HTML5/Theora, Safari and IE will get Theora support.
Nice thing is, Youtube wont change in at least a few years.
2) Rest of the web, will be able to push its prefered codec before youtube switches. Now, if we manage to make video usage more common in websites, and use HTML5/Theora for it, other browsers will have to follow.
Wikipedia’s new movement is exactly a huge step forward to it, but its not enough.
So, people who think web needs to be free, please, instead of youtube, use dailymotion. Push HTML5. Thats all you can do before youtube pulls the codec-wars trigger.
Mozilla is GPL licensed. Read the GPL and you will see that it explicitly prohibits royalty bearing code. Mozilla does not have a choice in this matter.
Their only alternative I see would be to create a separate plug-in containing the H.264 codec, but plug-ins are the very thing HTML5 is trying to get rid of. That’s essentially equivalent to using Flash for H.264.
The H.264 people are unlikely to give Mozilla a GPL compatible license to their patent. If they did, all of the embedded vendors would stop paying licenses and switch to the GPL’d Mozilla code base.
Mozilla is stuck. We need Google with their On2 acquisition to release a modern, free codec and get everyone out of this pickle. By switching YouTube to this new codec it will force all of the browsers into line.
no, it is MPL licensed.
Mozilla is tri-licensed – MPL, GPL, LGPL.
http://www.mozilla.org/MPL/
I have contributed code to Mozilla and had to sign the licensing agreements.
Trying to search for more infos on On2 VP8, I found this blog post from a x264 dev…
http://x264dev.multimedia.cx/?p=292
What strike me as I was reading it was the prospect that html5 video tag might just fail because of the codec issue, despite its promise to solve our “I hate Flash” issue.
That is COMPLETELY out of touch with reality. First of all, back in browser wars 1.0, NOBODY cared about standards. Back then, standards were based on the things the vendors built, not the other way around, and both sides were trying to compete based on features, not standards compliance. Back in the day you had CSS and you had LAYER (the netscape styling api), you had javascript and you had jscript. It wasn’t about IE independent, it was about supporting netscape, which itself was full of non standard proprietary extensions.
The only IE specific sites nowadays (outside of korea) are ancient webapps that companies don’t want to spend money on re writing. They are the equivalent of ancient timesheet systems that employees have to use terminal emulators to use.
Also, the web doesn’t rely on IE6, the issue is that 20% of the world is still using it, and that is too big a number for anything but hobbyist sites to ignore. If you have a business, you would have to be deranged to turn away 1 in 5 potential customers.
As it stands now, webkit is the browser of the mobile web, and ie has a 60% market share. Ogg will be something you have to do to support firefox, developers will hate it cause it makes their job harder, and companies will hate it because it doubles costs.
Actually you can technically browse the majority of Korean websites with Firefox. But the problems is that ActiveX and Adobe Flash is everywhere in the design components.
What the main issue is that ActiveX is everywhere in the Korean online banking systems.
Opera 10.50+ on Linux plays pretty well H.264, not only Theora. In fact it plays practically all possible file formats which is supported in gstreamer library (which beside own codecs can also use ffmpeg libs).
For example today i found Youtube to server HTML5 videos also to Opera. They are encodec in H.264, but they decided to let browser decide alone which codec it supports (and this is good decision). Browser will eventually automatically switch to fallback mechanisms if video with given encodings are not supported.
And this works greate. It make me mad to see pages which do some hardcoded browser detection, and assumes that i.e. Opera is not supporting something. Good because this is bullshit, or will be bullshit in the future. And only browser(and user) alone should decide here, not site developers.
Also Chromium on Linux is supporting other codecs than H.264 (on linux it uses not licensed libraries accuired by Google, but ffmpeg, which supports actually every imaginable video format)!
This problem (browser vs. video capability detection) is even more important in case of some less known browsers. There are few good browsers which even i don’t know names, but which have working HTML5 video with some combinations of supported codecs (some supports everything, some only H264 ,some only Theora). So most important is now to make webdevelopers know how to properly use HTML5 video. (preferably with no javascript at all, and with PROPER fallbacks, which are not based on browser Name/Version).
Edited 2010-03-19 17:10 UTC
…which is illegal in the US.
Agree. MPEG-LA claims to represent 900 patents(!) which are essential for an implementation of H.264. Fighting these in court based on obviousness or prior art would never work and the last one will expire in 2028. The only solutions are avoiding it and fighting for software patents to be abolished.
Related info I’ve gathered:
http://en.swpat.org/wiki/Campaigns_to_avoid_certain_patented_ideas
http://en.swpat.org/wiki/MPEG_video_formats
http://en.swpat.org/wiki/Audio-video_patents
http://en.swpat.org/wiki/Harm_to_standards
http://en.swpat.org/wiki/Use_software_from_20_years_ago
swpat.org is a publicly editable wiki, help welcome.
Edited 2010-03-19 17:09 UTC
stick with theora, but get as many other non-patented codecs in as well. dirac has just released a significantly faster decoder (the old complaint was that it was slow).
There are two implementations of Dirac. Dirac reference encoder, and fast Schrodinger encoder. I was using schrodinger but it was quite slow. I think this slowness was not becuase of codec implementation actually but more because I was testing huge HD video, which even encoded in H264 are not watchable on my Linux computer (1.7 GHz Pentium M with ATI M22).
I see Schroedinger-1.0.9 is released. Will hopefully test it soon. Also will test it in Opera and Chromium on Linux to see if they play well.
EDIT: OH, i already have schrodinger 1.0.9. Debian Unstable is so awsome. Going to test it now
Edited 2010-03-19 17:46 UTC
If Google ever chooses Theora, Microsoft could be the one to showstop the migration (real question is, why?). This could however be mitigated by either persuading Adobe to support Theora in Flash player or that Google builds their own video player plugin (or take gnash, fit Theora support and make it run on IE ,Safari etc.). Also their content is encoded in h264, so they have to keep the support in.
If Theora will be supported on each browser platform in one way or another, they can really _choose_ the future of internet video. We all hope it’s going to be Theora since MPEG LA could start asking for huge royalties in the future.
The img tag supports jpg, gif (once patent encumbered by Unisys), and png. Why not do the same for the video tag? Admittedly this will not solve the license costs for browsers, but at least the issue between choosing one or the other would be settled.
Thing is, IMG was supposed to be deprecated in the next-gen HTML in favor of OBJECT – which was supposed to be the ultimate in REAL freedom.
What do I mean real freedom? FORMAT AGNOSTIC. If the host system has a plugin, codec or decoder to handle the format, the OBJECT tag was supposed to use it – be it image, video, audio or a language applet!
But the whacko’s who’ve hijacked HTML5 adding all this extra unneccessary bull have completely backtracked on the simplification and freedom that the STRICT HTML4/XHTML1.0 doctypes started us down the road on. Now instead of making it simpler they just want to throw in two or three dozen new tags and attributes for god only knows what, instead of just riding microsoft’s ass about not implementing OBJECT properly (which they JUST did with IE8 for the first time)
VIDEO and AUDIO, as new tags for HTML is pure manure – as is two-thirds of this new ‘specification for nothing’ that is only going to result in an even more convoluted mess than the people still coding HTML4 like it was HTML 3.2 and this was 1997.
View Source on most websites to see what I mean – we’ve only barely broken people of tables for layout, and even when tables are appropriate most coders can’t seem to be bothered to learn how to use them (there are tags OTHER than TR and TD that go into tables) – so now they’re going to add more tags for people to not learn properly and abuse by nesting until blue in the face? BRILLIANT!
Hey, an XHTML2 troll. You don’t see many of them these days.
As others have pointed out, there is an extremely simple solution to this problem: all Mozilla has to do is make the codecs supported by HTML 5 installable as plugins and only ship the free Theora plugin as the default. Rather than forcing their decision on their users just like the other browser vendors, Mozilla could let their users choose to stick with the default Theora codec or install a H.264 plugin from a third party if they choose.
Only the first fix is free.
Mozilla is open source. Deliver Mozilla/264 if you care so strongly about it.
Maybe that isn’t even necessary, as the extension interface is good enough?
See, two ideas how to approach this without Mozilla effectively endorsing non-free codecs.
Gif could be freely decoded, just not freely encoded. http://directory.fsf.org/project/libungif/
H.264 can neither be freely decoded or encoded.
Because not supporting it if it’s already present on the host system gives users it… NOT
You know, this crybaby “Use Theora” bull is wearing JUST a little thin. Maybe if theora grew the *** up and had REAL hardware support, REAL corporate support, and a practical business model behind it you might see legitimate support.
Without any of that, it’s naive pipedream bullshit and little more than philisophical grandstanding one step removed from the typical open source rhetoric spewed forth by the back room unix server geeks who were left behind by the real computer revolution and echoed by the young and gullible college age techno-hippies (let’s call them Tippies) who don’t understand just what’s so wrong with communism.
Said computer revolution being driven by CORPORATIONS providing the financial backing and muscle to get the products people want into their hands. Corporations like Apple, Microsoft, IBM, Borland, Lotus, Tandy, DEC, Novell – and while some of those have fallen by the wayside over the years, without those Corporations and the people who work at them *nix SH would probably still be considered a bleeding edge UI.
But rah, rah, fight the power – keep on throwing out the word freedom without understanding what it means… Or grasping that joe sixpack could give a flying **** about your ‘free as in freedom’ and is more than willing to pay a real company to produce a real product that real people might want to actually use.
Lands sake, grow the *** up already!
Edited 2010-03-19 17:57 UTC
My, you need a lot of expletives to get your point across, don’t you?
So, uhm, what was the point of your comment? Communism is bad? Corporations are fluffy bunnies we should hug and kiss?
What is it? This time, try it without sounding like a 12 year old who just got his first dictionary!
To quote George: “When I want my men to remember something important, to really make it stick, I give it to them double dirty. It may not sound nice to some bunch of little old ladies at an afternoon tea party, but it helps my soldiers to remember. You can’t run an army without profanity; and it has to be eloquent profanity. An army without profanity couldn’t fight it’s way out of a piss-soaked paper bag.
As for the types of comments I make… Sometimes I just, by God, get carried away with my own eloquence.”
You want the short version?
The ONLY reason to choose Theora over h.264 is the ‘free as in freedom’ rheotoric, which is more fiction than fact – and not enough to offset that companies WANT to support h.264, we have hardware already in the market going back half a decade that can do hardware decode on h.264 (there are Ge6’s that can handle it) – and the majority of users out there are running operating systems that come with the decoder.
Naive idealism and dirty-hippy pipedreams about some fictional ‘freedom’ that is in fact the exact opposite of freedom is not a legitimate reason to choose a codec! This is why REAL companies like Google, Apple, Adobe, HTC, Nokia and Microsoft – even when said companies that don’t agree on HOW to implement their video and are even in the middle of all suing each-other like they hired on ex-SCO lawyers – all have little to give Theora except their middle finger.
Edited 2010-03-19 18:20 UTC
I really don’t like being called a hippie. You may think acting like a jerk makes you cool, and you can hide behind your “I-get-carried-away” nonsense, but if you can’t present your argument without straight-up insulting me then I really see no point in listening to your arguments.
Maybe this is the way you do things is where you live, but I’ve been raised properly, and this nonsense is beneath me. I’d rather be arrogant than a jerk.
That’s high comedy given your track record on some subjects.
I’m not so sure about that. Note:
Which is actually a self-consistent view, if not perhaps noble. And do note, he manages to get his point across without the use of comically-asterisked-out obscenities, or without his posts being fifty-percent ad hominim attack by length. With the lack of substance being covered by comical frothing-at-the-mouth yelling and name-calling, your posts would be perfect satire of your own view, if I wasn’t fairly confident that you really mean it.
The question stands: if you can’t carry on a conversation like an adult, why should the adults listen to you?
Edited 2010-03-22 23:41 UTC
So where’s the difference between inferring someone is a hippy and labelling someone “fanboy”?
Considering how much you rant about software patents yet still work for a site that uses an adnet that is patented up down and sideways.
Guess using software patents to make money is OK as long as you and yours are the one making a profit.
Go back to masturbating on Ayn Rand (or George Patton), moron. And when you’re done, grow up.
Nothing you say makes any sense whatsoever. You’re just regurgitating extremist right wing rhetoric.
Sorry bud, but I’m the customer, and the corporation/business/whatever are beholden to me, else I take my business elsewhere. I will never let myself be beholden to the corporation. They have to earn my business. That’s the way free enterprise, capitalism, and market driven economics works.
Ogg Theora works fine, and doesn’t lock me into the wares of big corporations. h.264 works fine too, but it does lock me into big corporations.
You are clearly too stupid and ignorant to see how h.264 is a tool for lock-in, and stuffing out smaller and/or open competition.
And why go from one proprietary format (Flash) to another even more proprietary (and more expensive and more restrictive) format (h.264), in the guise of of open, standardized HTML5? That makes no sense whatsoever, and it’s far, far, far from pragmatic.
And in case you’re wondering, I work for a proprietary software company. I’m all for proprietary for some things. But I’m against unreasonable restrictions, unreasonable license fees, and unreasonable lock-in.
That’s not hippy BS, that’s pure pragmatism, as well as pro free enterprise.
Because it’s so unreasonable that Adobe cannot effectively give away the decoder to over 90% of the people who watch videos online and couldn’t possibly find the handful of pennies for including the encoder with CS4 – and certainly companies with minimal revenue streams almost entirely based on advertising, like say… Youtube… cannot possibly afford to distribute videos in that format.
Oh wait: “Formerly limited only to over-the-air broadcasting, ‘Free Television AVC Video’ now refers to AVC Video that constitutes television broadcasting which is sent by an over-the-air, satellite and/or cable Transmission,” – so much of this ‘free as in freedom’ licensing bull does not apply if you aren’t CHARGING for the video under the MPEG-LA license… and anyone who’s putting this stuff up for profit, can probably afford pennies on the dollar JUST like DVD and Blu-ray vendors do with their releases… It’s a NON-ISSUE!!!
Right, unreasonable licensing. That’s what h.264 has, sure it is.
In the interest of community, Pete’s group have pitched in together and colaboratively developed a competitive codec that anyone may implement and use anytime, anywhere, on any platform, in any context. Pete’s group charges no royalties at all, and the license they use guarantees that no royalties will ever be applied.
Mike’s commercial group is charging royalties (even though they are “pennies on the dollar”) for their equally capable codec such that individuals and small groups, and even some buisnesses in developing nations, cannot afford it. The original development cost of that codec to the commercial companies in Mike’s group has been recovered, many times over, years ago, so this ongoing take now is pure pure cream. Mike’s group gives permission to use their codec only to select paying groups that are allowed to be “in the club”, apparently in the interests of reducing competition against themselves. The basis via which Mike’s group is even allowed to charge royalties is very dubious. Supporters of Mike’s group (crazy, I know, but there are some) spend an extraordinary amount of time and effort disparaging Pete’s group.
For myself, I’m opting to go with the codec from Pete’s group. I’m going to support that one to the hilt, because after all, it does follow the whole “access for all” tennet of the web in the first place.
Edited 2010-03-21 22:45 UTC
Are customers really liable for the there companies patent licences. Imaging having bought a (real hardware; easier to imagine) video player that uses the H264 codec from the company Videovistavision . Are you really liable if they did not pay any licence fees?
I really can’t believe this. That means that you’ll have to study each company from who you consume, which you can’t even if the hand out there administrations to stangers.
And yet, it is reality. MPEG-LA themselves have clearly said so.
Since when did what the MPEG-LA constitute a legally binding judgement? Only the courts can say if customers can be held liable. Contracts / Companies will claim all sorts of crap that isn’t true in the hope of scaring people into believing them.
Yeah, because ordinary folk stand a real chance against the MPEG-LA. In world where sharing a few songs online can lead to hundreds of thousands of dollars in damages, do you really think the courts will hesitate to do the same to people who make an H264 video or two but don’t have a license?
Naive, naive.
The ridiculous damages for copyright infringement are applied because that is what the law states is applicable, not because the RIAA say it should be so. No naivety whatsoever.
Since this:
Can in many ways be considered revisionist history. It is very easy today to forget why IE6 and IE 5 before it gained such massive market share. Fact is, for all the picking on IE we do today it was in fact the most advanced and standards compliant when they were released.
Netscape had Microsoft on the ropes – in 1998 netscape still had over 55% market share by most numbers – IE 4 wasn’t even making a dent with it’s 30-40% share… and then something happened.
Microsoft released IE5. It fully implemented HTML 2 according to the specifications – it fully implmented HTML 3.2 according to the specifications – and even implemented the not even out of draft CSS. In fact, the broken box models and other ‘issues’ we have supporting IE using ‘standards compliant’ code comes ENTIRELY from people using the draft implementation on production sites before it was finalized – and people wonder why I rage against using CSS3 on production sites? IT’S NOT OUT OF DRAFT PEOPLE!!! Be fun a decade from now when we’re talking about gecko/webkit backwards support for the draft specification and not being able to use the ‘proper’ final model.
To top if off Microsoft went and took all the stuff that was netscape only and copied it so all those ‘best viewed in netscape’ pages suddenly worked fine in IE – and even more so added a lot of scripting functionality that Netscape could only have dreamed of and that today are commonplace.
AJAX? IE’s had that since Office 97 came out – XML requests are a Microsoft innovation for pete’s sake. ActiveX gave developers control over the browser (did we mention that from a purely functionality standpoint there is little difference between and activex control and a firefox plugin?) – and what happened?
Something we make fun of Steve for, when in fact he hit the damned nail on the head; DEVELOPERS, DEVELOPERS, DEVELOPERS, DEVELOPERS, DEVELOPERS, DEVELOPERS.
Developers who were around BEFORE 2001 didn’t flock to IE6 because it was the defacto standard – they flocked to it because IE 5.5 and newer offered development features that other browsers didn’t offer at the time. Firefox wasn’t even a twinkle in a FLOSS fanboy’s eye, Mozilla Suite was a buggy bloated train wreck that only fringe whacko’s even TRIED to use, Opera didn’t even have CSS support yet, Netscape was effectively defunct with every release after 4.5 being little more than bugfixes, and it gets even uglier from there.
Developers flocked to IE 5.5 for what it could do that Netscape, the only real competitor, could not. You saw more and more ‘best viewed in IE’ and less and less “best viewed in Netscape” – until by the time IE6 hit, gave us a bug fix to be more standards compliant that 5.5 was (doctype trigger)… It was the best documented most feature rich browser – nobody else of it’s time was even CLOSE.
The only reason it became a problem is that Microsoft then sat there with their thumb firmly up their tuchas for five years and dissolved their entire development team.
I wouldn’t call it pragmatism, I’d call it market reality… and the reality also is that once Firefox turned gecko into something USEFUL and mozilla cleaned up gecko’s act to be more standards compliant (We often forget what a train wreck Gecko prior to 1.0.1 was on ‘standards’ since it still behaved pretty much like NS4 – hell, who are we kidding, it still does on a lot of ****) and all those things IE could do other browsers couldn’t were added it was embraced by developers.
That the indoctrinal rheotoric of the open source community cultivates the young into joining the cult doesn’t hurt on the developer front for it either…
But to say developers went to IE6 just because of market share is shortsighted and foolish, since it forgets why it got that share in the first place. It only took four years for IE to go from 15% use on IE4 to dominating the market on IE 6 – you make those types of radical in-roads by making products people want to use and developers want to develop FOR. See firefox and how it’s rapid climb (which I suspect is artificially inflated thanks to superfetch BTW) in share has been over a relatively short period as well… Even 0 to 30% of an established market in six years of public releases is an astounding feat in the business world.
You then go on to belittle web developers for choosing flash – lemme ask you this; Cross platform, cross browser, what choice was there a decade ago? Realnetworks? **** please… I know the euro-fetish crowd have always had a raging chodo to get their spanking porn via REAL, but everyone else on the planet hated REAL’s guts. You couldn’t rely on OBJECT or the proprietary EMBED to load the host codec via WMV/AVI/MPG, which didn’t even OFFER controls – Adobe offerred a simple, professional product, that included development software for editing, creating, and doing a whole bunch of things that nothing else on the market came CLOSE to offering.
Taking the developers to task for originally choosing flash is like taking the developers of Quake II to task for only supporting 3dfx minigl out of the box – what other choice was there… That developers are sloth to even CONSIDER switching away from a technology that has served them well for over a decade; This is a suprise to you?
Lemme ask you this: Is there an official suite for handling theora based video using the HTML5 tag, SVG and the associated javascript along with action-scripting sound that’s as complete as what Macromedia offered with Flash five years ago when Adobe bought them out? Hell, flash 7 is still more robust than any of the open specifications that have even been implmented when it comes to making the developers life easier.
We really need to stop making fun of Steve when he’s right.
Though it feels a bit wierd defending Adobe given my opinion of Dreamweaver.
IE5 fully implemented HTML 2? I don’t think so. (and so the same applies to 3.2)
Where was the full support for link for example? Something we still lack in most browsers. Mozilla implemented it giving us a browser navigation bar for up next previous etc. which link supports but I don’t think Firefox does (No an extension does not count as supporting it if one exists)
I thought EMBED was a Netscape supported tag, not IE (I may be wrong on this one) and it existed in the HTML 2.1e IETF dtd.
Parts of CSS3 are out of draft as CSS3 consists of a number of modules with different paths to recommendations.
You couldn’t rely on it partly because IE has only just got a non broken implementation of object and developers decided to force a specific ActiveX control for playback instead of just saying here is an object in this format. Also as stated earlier EMBED was in an IETF dtd so not proprietary.
According to Wikipedia, Mac IE 5 was the first major browser to offer full (~99%) CSS1 support.
http://en.wikipedia.org/wiki/Css#Difficulty_with_adoption
http://snipurl.com/uy058
Edited 2010-03-19 23:54 UTC
Half right on both our parts. It was a netscape tag originally, IE added support to it in 5.x as part of their trying to make the ‘best viewed in netscape’ thing pointless…
Though using an IETF dtd as an example of ‘part of the specifications’ is questionable at best since the IETF does not define the official HTML specifications, the W3C does. The Internet Engineering Task Force’s job is the TCP/IP specification, not the data that goes across it.
That’s a bit like saying the department of public works has a say in the diameter of a car tire.
There are a lot of ‘secondary’ DTD’s released by all sorts of organizations as a form of HTML/SGML/XML that have nothing to do with the official HTML specifications set by the W3C.
What’s the old joke, any idiot can type up a whitepaper and call it a specification.
Edited 2010-03-21 20:05 UTC
The image tag supports PNG and GIF, the video tag should support h264 and Theora. Problem solved.
h264 is a step in the right direction, Flash is more propietary, more patent encumbered. h264 is not perfect, It’s not Free but It’s better.
The image tag does NOT support ANY specific image format. PNG, GIF, and JPEG support are simply defacto standards. There is not and has never been a requirement to support any of these formats for inline images in any W3C standard – it would simply be stupid for any browser not to support them because of their widespread use. Hopefully at some point it will become stupid enough not to support Theora that everyone supports it – that is the best outcome that can be realistically expected now. The standardization process is essentially over – that ship has sailed.
H264 becoming the defacto standard on the internet is the more likely outcome at this point – it is also the most dangerous and unlike the whole GIF fiasco the patents and licensing on h264 are being actively maintained and there will almost certainly be a money grab at some point – anyone thinking otherwise is simply clueless. It may take years before the other shoe drops but having a heavily patent encumbered video standard is a BAD idea…
So if H264 is where this whole thing ends up landing so be it, but I personally am going to keep up with Theora as much as possible so when the shit hits the fan I can be ready to re-encode everything if need be. This isn’t how I would have wanted things to end up – I like a lot of other web developers would much rather Theora support be required NOW. But with only Mozilla squarely supporting it – I’m afraid at this point Theora is probably doomed. Eventually MPEG-LA will go too far and instead of everyone switching from GIF to PNG, we will be switching from H264 to Theora. That is how the world turns – I’m used to it by this point…
Just to summarize – when the rubber hits the road and the patent encumbered nature of H264 rears its ugly head (and it will) there will be a rebellion and the rebels will win. Everyone will forget how great H264 is when the reality of having to pay for the privilege of using it hits them…
GIF could be decoded without the LZW patent, which was only used for encoding.
I like how some people’s idea of freedom and choice is based on telling people what’s good for them and limiting their choice.
No, I’m not happy about the H.264 license situation. But trying to push Theora down people’s throats makes Mozilla no better than all the corporations trying to push their favorite solutions on to people.
It’s about “freedom”? Then give people freedom. Make it a plugin system. Ship it with only a Theora plugin if you will. But give people the freedom to add support for the format they want.
Everything else is not “freedom”.
How are they pushing theora down people’s throats? They’ve just chosen to not support h.264 out of the box for the HTML5 video tag. Guess what – Firefox also doesn’t support Flash out of the box – it’s a plugin … same with WAV, same with Quicktime.
They’re pushing it by not allowing integration of other formats.
Flash and Quicktime support are good examples for what I wrote in my parent comment: extensibility through a plugin interface.
It annoys me how Mozilla’s stance is characterised as ‘idealistic’ against the ‘realists’ and ‘pragmatists’ of the H.264 camp. In reality, that ‘idealism’ is simply a pragmatism with a view to the future slightly longer than the next 6 months. An internet where the basic tools to create, view and edit content are under the domain of a single profit-motivated entity is an internet that progressively bars those without bags of money from participating.
Moreover, why is noone raising the question that the two holdouts – Apple and MS – are both part of the MPEG-LA patent portfolio?
(And Gruber is a douche and an apple fanboy – willing to embrace openness over flash in one moment, and in the next call for prgramatism over H.264).
I’ve read this whole thread and I didn’t understand most of it. I’m not a native english speaker, so my english skills might be failing me here, but I didn’t know that “pragmatic” and “short-sighted” were synonyms.
So yes, using h.264 looks like a very “pragmatic” decision.
This sentence in the teaser, and a lot of what was written in the article, frankly pissed me off… I am a LONG time web developer (about 15 years) and I firmly believe that for the video tag to be a truly useful and relevant part of HTML5, it should (ideally):
1. Have a default codec.
2. Support for that codec should be required in conforming user agents.
3. Said codec should conform to the W3C RF Licensing Policy (without requiring any special exemptions).
I certainly know a lot of other web developers who feel exactly the same way – I do not think your notion of how web developers feel about this issue is in any way accurate.
Inferring that web developers on the whole are backing h264 just because of the ramblings of two attention whoring bloggers is a bit of a stretch imho. It is not web developers that are pushing for h264 – it is content publishers and the companies backing _some_ of the major browsers.
I have nothing against h264 – in fact I will completely admit that on the technical merits alone it is a marginally better codec for streaming purposes and frankly a much better codec for low/high bitrate file encoding (imo in the middle of the spectrum the differences are not as pronounced). In most cases it generally produces better quality at the same bitrate, it has a much more developed tool chain, its more flexible because of its more refined complexity targeting, etc. etc. I think those arguing for Theora on the technical merits are just distracting from the real argument – because if h264 was NOT patent encumbered we would not be having this discussion at all…
The point is right now, Theora is the best choice for a default codec, and for no other reason than that it is NOT patent encumbered. That’s it – there is no other argument needed for its merit. Even if it was MUCH worse technically than it actually is – it would STILL be the best choice. The default choice does not need to be the best – it simply needs to be uniformly supported and free for both producers and consumers to use without any strings attached. Theora IS the best choice out of the available choices, because h264 simply doesn’t qualify as a choice.
If Theora (or some other acceptable alternative) were the default codec, and all the browsers decided to ALSO support h264, I have no doubts that realistically Theora would never catch on – it would be eclipsed in usage and it would be relegated to use only by “non-commercial” users. You know what? That is perfectly acceptable to me – because I don’t care about Theora – I care about being able to produce a video and publish it without any licensing requirements are legal intricacies. The point is that since such a codec exists it SHOULD be made default.
ps. The reason web developers backed IE for so long was that at the time of its reign there were no real standards to speak of. Have you ever actually read the HTML 3.2 standard? It is very simple – there frankly isn’t much to it. And frankly IE was the most conforming of all the available browsers at the time. Proprietary extensions in browsers were the norm – not the exception. Most of what you see as standards today started out as proprietary extensions by either Netscape or Microsoft – That VAST majority of the standard W3C DOM model is based on Microsoft’s proprietary work. Their model was simply better than Netscape’s for the most part (document.all and such atrocities not withstanding).
It is totally wrong headed to blame web developers for backing IE at the time – that is how things became standardized in the first place. When the standards process finally caught up to reality, and it became possible to actually put the horse back in front of the cart, web developers rejoiced. They were some of the most outspoken advocates for the process in the first place. That’s just my 2 cents…
I gotta admit that was where a lot of my rage came from over this one too…
I mean, if you’re not going to listen to the people working in the field, who understand the technologies involved and have a motivated interest in delivering the content of their sites to people in the best way possible – who are you supposed to listen to? Some washed up emacs programmer who probably doesn’t realize you shouldn’t have more than one H1, that fixed metric fonts are an accessibility /FAIL/, that IE conditional comments are a waste of bandwidth, etc, etc?
That’s like saying don’t listen to an architect when building a skyscraper or don’t listen to a gunsmith about maintaining a firearm!
Like much of the “free as in freedom” rhetoric being spouted to promote the OPPOSITE of freedom, people are dumber for having read things like that.
Edited 2010-03-21 18:08 UTC
How many normal, every day, non-geeky users do you think even know what Theora or H.264 is let alone care which one is used for web video? How many people still have no idea that Internet Explorer is just one browser option?
While it’s ALWAYS better to use opensource components where possible – and PLEASE stop referring to Theora as a “standard” because it is NOT – the reality is the choice will be driven by the browser market, and between IE, Safari and Chrome there’s a sizeable chunk of the market taken. Do you honestly think users, many of whom don’t even know they’re using a “browser”, they just click the big blue E to “get the internet” – will change browsers?
How many of the existing video rental or download services use Theora? Do you truly believe they will want to encode on one format for one medium and a different format for another? We already know some of the shortcomings of the Ogg container so it simply doesn’t lend itself to commercial applications, which creates the issue of encoding in multiple formats.
This argument can go on ad infinitum in groups like this but it won’t change jack in the real world. H.264 is the only currently recognised standard, and until that changes Theora won’t have a look in. I’m not saying we have to like it, but that’s how it is. There’re many things in this world that we don’t like, that list varies for each of us, but much of what’s on there in every case are things we can do nothing about. Right at this point in time this is one of those things.
If you want to change it get the RECOGNISED standard changed by putting up your money like those who got H.264 into that position did. Because at the end of the day it IS all about the money. Again, not saying that’s how it should be, but that’s how it is…
blaiming => blaming
moon => Moon
How about shunning vendor-specific technology, like a browser’s video support limiting you to certain codecs?
You know what’s awesome? Xine.
Also awesome is Gstreamer.
Not quite as awesome, but still pretty awesome, is Directshow.
By forcing Theora, the cure is becoming almost as bad as the disease (In MS’s or Apple’s case, I’d say worse, but Adobe is right up there with Creative).
Firefox goes to great lengths to integrate itself on its popular platforms. Do that properly with the video tag, please, so that I can view an H.264 movie, or a Theora one, if I want, using my already-installed codecs, through the video tag. It can’t be that hard.
Edited 2010-03-20 02:03 UTC
This is not an option. Different operating systems come with completely different sets of codecs.
For example, Mac OS X comes with support for h.264 in an MPEG-4 container, with either MP3, AAC or AC-3 audio. Fine.
Earlier versions of Quicktime, or the versions included on the Apple TV, iPhone, iPod Touch and iPad, all include different levels of h.264 support. It wasn’t long ago that even Quicktime on Mac OS X proper didn’t support anything beyond the h.264 baseline profile.
Windows 7 includes support for h.264 in an MPEG-4 container, with MP3 or AAC audio. It has a full h.264 decoder, so that’s OK.
Neither Windows XP nor Windows Vista have this codec. Instead, you get access to Microsoft’s proprietary codecs (WMV9, WMV8, WMV7, WMA1, WMA2, and the ASF container format), and nothing else.
There are no free, legal h.264 decoders available for Windows XP or Windows Vista. At all. So, if you wanted to view h.264 video on one of those platforms, you’d either have to violate all the patents, or buy a decoder.
As for Linux… GStreamer doesn’t come with a standard set of codecs. You can probably rely on most distros to include Ogg, Vorbis and Theora, but little else.
Legally licensed decoders for various patented formats do exist. Except for the MP3 one, you have to buy them.
http://www.fluendo.com/shop/category/end-user-products/
You’d have to pay 40 Euros to get a legally licensed h.264 decoder for Linux. Same problem applies as for Windows XP and Vista – there’s no way in hell you’re going to convince people to buy a codec.
What do you notice about those platforms? In the baseline codecs, there is not a single, common codec among the bunch. If all browsers simply used the system codecs, you’d be in a situation where you need to include THREE versions of each video – h.264, Theora, and WMV.
If that persists, then obviously, H264 by itself would lose, without being directly snubbed by a browser vendor, else some major licensing changes would come in, and Ogg Theora would win, being freely distributable.
Clearly, the patents are broken, and/or licensing is broken, and/or the OSes and their maintainers (including MS and Apple) are broken. Copying their restrictive behavior, though, would make Mozilla equally as broken, despite it being in the name of freedom.
I also see no problem, with, say, Firefox going native on each platform, and on the two where they do the bundling (Windows and OS X), including Ogg Vorbis and Theora decoding support, provided that it either adds a system filter,or at least has a fallback to the system’s own native capabilities. Then, also make them required dependencies for *n*x systems. Just don’t make them the only option.
Edited 2010-03-20 09:54 UTC
It is an option. In fact it is the only sane option. Use the media framework of the system.
You do have a point however and the solution is simple. The browser should make sure the codec is installed, installing it if necessary.
I don’t really have a problem with Theroa being the recommended codec just so long as the implementation is done so in a sane fashion and not hardcoded and built into the browser disallowing the option to use something else. Going this route would have an added benefit that the codec would magically work with other tools on the system so people could create Theroa videos in their app of choice.
While they claim they are, it has yet been prove.
The main difference is that for h.264, there are several parties claiming 900 applicable patents (incl. Apple and Microsoft, which is a certain conflict of interests)
For Theora, there are 0 parties claiming 0 applicable patents so far.
What does that tell us?
1. Theora might be free from patents, h.264 isn’t free from patents.
2. For both formats, there might be some additional party claiming to have applicable patents. I haven’t seen a public statement by MPEG-LA that they indemnify licensees against patents of _other_ patent holders, and their licensing contracts aren’t readily available.
What’s the risk?
With h.264: That 2015 they cash in on their 900 patents and that some additional party might pop up (maybe tomorrow) and demand money
With Theora: that someone appears with claims due to a patent (who didn’t want to cash in on the patent with On2 already)
So they share a common risk, while h.264 has an additional risk.
I think you’re barking up the wrong tree. Mozilla has zero power to influence the choice of codec. The decision will be driven primarily by content owners/producers (eg. YouTube, Hulu, etc) and OS platform vendors (Apple, Microsoft, Google, etc). This is ironic, given the fact that it should be the W3C’s role to endorse a particular video standard. But that didn’t happen. There are too many questions about technology, licensing, DRM, etc.
What I would suggest is that browsers need to insulate themselves from whatever standard emerges by adopting a standard pluggable model for integrating codecs that is agnostic to the choice of codec. This is going to be more important over time, because video technology is constantly evolving. H.264, Theora, and the current raft of codecs will be replaced within the next few years with something better. Unless there is a pluggable model for handling video, we will be back in the same position, squabbling over how to handle the change, licensing, etc. My suggestion would be to choose the appropriate level of abstraction.
As for Mozilla, there is another option. Google could donate the license fee for H.264 to Mozilla, along with the money that they give them every year for making Google the default browser search provider. Longterm, though, it isn’t in Google’s interest to keep Mozilla around. Google would like everyone to move to Chrome. Can you think of a better way to do that than by NOT solving Mozilla’s video problem?
Edited 2010-03-20 07:05 UTC
People will install whatever it takes to watch YouTube.
Web publishers will use whatever the majority has. You can’t expect small publishers to encode in multiple formats.
This means the company in the best position to rid the web of Flash is Google and they have in fact moved closer to Adobe in the past few years.
Those bloggers are right in that if Mozilla doesn’t support h.264 it will just keep Flash around even longer.
Opera’s refusal to license H264 hasn’t got anything to do with ideals, obviously (Opera being proprietary and all), but the Norwegian browser makers believes the license is simply too expensive.
That’s a new low for OSAlert. Well done Thom for spouting such utter and absolute tripe!
Opera supports and has always pushed for open standards. Man, your comment is so ill-informed. How the hell did you end up being so heavily involved in this site with that level of either childish nose-thumbing or inaccuracy? And no, I don’t know which is worse.
I’m just taking my information STRAIGHT from Opera. They have stated that they do not include H264 because it’s too expensive.
I’m sorry of that doesn’t fit within your view of Opera, but in that case, complain to Opera, not to me for just repeating THEIR words.
Really? You’re quoting their words? Go on then. QUOTE.
You’re full of crap because I can give you several places where they talk about the dangers of H264 to the web.
Who cares about all this?
Give up on this subject guys. It^aEURTMs just computers.
If “computers” is a subject we should give up on, and not care about, why have you bothered to make a profile and post a comment on a site that is all about computers?
How many friggin times authors on this site keeps saying “luckely I have FlashBlock”? I don’t care! Install whatever you want.
HTML5 is never going to replace Flash. Never.
I understand that for layout and “general websites” you are fed up when you see that they are created with Flash. I agree, that sucks.
But you cannot create cool interactive experiences like product campaigns with html5. You can’t create decent games with Html5. So if you block Flash, fine with me, but you block much more than annoying ads.
The answer is in your article: it’s about the devs, not about the technology. Flash 10.1 improved a lot (less CPU usage, more performance) but it’s up to the developers to use i
Flash where it makes sense.
My company specialises in the creation of webgames (Unity3d and Flash, both propierty tech) and I know that millions of customers are happy because of these plugins.
It’s just that you are a nerd. You are not the general public. Sorry to get ‘personal’.
Actually, the difference is in the dev tools. HTML 5 can do everything you mentioned, but the quick and easy design tools to do it the way most Flash devs expect to don’t exist. Until they do, Flash will still be around. Of course, given Flash’s inaccessibility, I will still continue to block any rich apps done in Flash.
The more the language used in this bothers me; “Shun H.264” – in other words don’t even support it…
Free as in freedom, so let’s remove people’s ability to choose what format they want to use… RIGHT.
Nowhere have I seen the false facade of this entire ‘licensing is evil’ rhetoric so crumbled and apparent as that right there. The discussion; the REAL discussion isn’t even about what formats should be supported, but what FORMAT should be supported ONLY…
… and that’s the key. Are we talking about making an API so all browsers could share codecs? Are we talking future planning so that if a better codec comes along or the market shows some other format is more popular the FORMAT maker can instantly add support for all browsers?
NO, people like Thom are talking about hard coding support into the browser for one format and one format only. Sure they’ll hide behind things like security (in which case having it hardcoded to the codebase instead of as a sandboxed plugin…) but really it’s little more than trying to restrict everyone else’s freedom to choose to their personal choice (theora) and their choice only.
Basically, much like Opera’s little hissy fits over the browser ballot or RealNetworks pissing and moaning about WMP, or the US Congress shoving HD down the public’s throat whether it wanted to pay for it or not, it’s the same “We can’t compete on our merits, so how can we force people into accepting it our way or nothing”…
… and that’s NOT freedom, and anyone trying to tell you otherwise doesn’t know what freedom is.
Edited 2010-03-21 23:22 UTC
If you are able to nominate another competitive choice for a web video codec that : (1) has no royalties; (2) incurs no liabilities or risks (particularly for individual end users) for its use; (3) may be used without restrictions by anyone (including individuals and nonprofits) of any nationality at any time on any platform for any purpose; (4) is not constrained to any platform or small subset of web access devices; (5) is available and stable right now, today; (6) has patent coverage; (7) has no patent claims against it; and (8) does as well as any competing proprietary format; then please, by all means, nominate it. We are all ears.
If you have no viable alternative meeting these criteria that is therefore suitable for use as the standard web video codec, then the one solitary candidate that does meet the criteria may as well be mandated as the web video codec, since having such a single standard makes implementation far easier (and also, as you say, more secure).
There is after all only one choice mandated for HTML, CSS, ECMAScript, SVG, DOM et al, and only one choice (TCP/IP) for the transport layers, so why should the video codec be any different?
Edited 2010-03-22 01:22 UTC
… and if 1, 3, 6 and 7 meant a damned thing to any COMPANY dealing with video, you might see progress; But they don’t. Those only mean something to the die hard fight the power anti-corporate FLOSS zealots, which is why businesses continue to look upon them as something of a joke.
In a number of cases… and this is something we’ve not really talked about – HTML5 video and theora are the opposite of what the majority of online video providers WANT; An open format with no way to mask the video download. Does the HTML 5 video spec support anything along the lines of RTMP and/or RTSP? No, it still does all it’s work via HTTP meaning that a certain industry is VERY unlikely to embrace it.
… and much like Betamax, if the porn industry decides it likes the rival technology better, even a quality difference wouldn’t matter.
Sad as it is, that is the part of the industry most likely to decide this for us – regardless of what’s “best”. There’s a reason they love flash based players wrapping a RTMP stream regardless of codec; It’s harder for people to copy without actually intercepting the data-stream.
Let’s face it, for ‘doorlock’ security (aka keeping the honest people honest) HTML5’s got bupkis.
Hmm. That comparison of Beta to VHS is a good one – it will come down to what the content providers want to use; and who’s the biggest slice of the content providers? An industry fighting tooth and nail to stop people from copying their content illegally.
Edited 2010-03-22 07:47 UTC
Companies don’t watch online videos. They aren’t the market.
A good deal of online video is user-submitted.
Nothing to do with Theora vs H264.
Given the matrices on this page:
http://en.wikipedia.org/wiki/Comparison_of_streaming_media_systems#…
it would appear that the Darwin Media Server has the widest support of operating systems, container formats and protocols. But this has absolutely nothing to do with codecs.
Hardly. The people decided which of Beta or VHS would win, not the industry.
A bonus for the people. It is a feature, not a deficiency.
There is a market of people. There is an industry of suppliers. The biggest slice will go to whichever has the biggest intersection. The industry might be all-fired keen to supply DRM-laden content delivered by royalty-bearing software, but it won’t fly at all if the people opt for participation instead.
Having the people actually participate (Facebook, Twitter, YouTube, Myspace, email, interactive games, even P2P) is what is driving success on the web right now. There is no market yet for huge DRM-laden company-sponsored video.
If h264 corners the market for web video, and big companies breathe a sigh of relief and start charging all and sundry, apply DRM and suing end users, that will kill it. Given the current brouhaha, it might already be killed.
PS:
Companies are not the web. The web is not designed for companies. It is designed for, and intended purely for, people to use.
http://en.wikipedia.org/wiki/Internet#History
The web goes back to the late 50s. The opening of the network to commercial interests began in 1988. Johnny-come-latelys.
Edited 2010-03-22 08:44 UTC
Backup:
http://www.youtube.com/
The logo says “broadcast yourself”.
http://www.youtube.com/t/about
http://videoonwikipedia.org/index.html
http://openvideo.dailymotion.com/en/about
http://www.vimeo.com/
Edited 2010-03-22 10:42 UTC
To websites run by companies looking for the best way to let users DO THAT while affording to pay for the HOSTING and technologies to do that!
The page also says YouTube LLC in the bottom corner. Limited Liability COMPANY
Which I’ll be honest, I never even HEARD OF until you linked to it, however…
It’s backed by the “Wikimedia Foundation” which if you investigate further…
http://wikimediafoundation.org/wiki/Home
Is incorporated with a board of trustees. In other words a company – the first I’ve seen supporting free codecs but being a NPO entirely funded by donations it makes sense for them…
But pay attention, I wasn’t saying don’t use theora or don’t support theora – I was saying there should be NO DEFAULT and all formats people WANT to use should be supported by the user agent.
On to DailyMotion, back at you with a link
http://openvideo.dailymotion.com/en/about
They’re soci~A(c)t~A(c) anonyme, the French Equivalent of a LLC. That’s called a business… company… corporation even. Business plan, money-flow, ability to have an income to pay hosting costs.
http://www.vimeo.com/
TM + ^A(c)2010 Vimeo, LLC
Another LLC – do I see a trend here.
I do find it interesting that all but ONE of the places you linked to have no interest in Theora – and all but one of the places you linked to actively uses scripting tricks and/or secure channels to try and prevent people from copying the videos off the websites…
It’s why YouTube supporting HTML5 doesn’t make a whole lot of sense given their longstanding practice of trying to prevent ‘their’ content from being copied to other sites without using THEIR interface for doing it. That’s one hell of a policy back-pedal.
People use the internet; Companies are what keeps it afloat. Companies like… OSAlert LLC?
Edited 2010-03-22 16:31 UTC
openvideo.dailymotion.com and wikipedia both use HTML5/Theora. Thats two out of four for those of us who can count.
The “interface” as you call it has absolutely nothing to do with the codec.
Vimeo and Youtube are latecomers to the HTML5 scene, and so far they have simply offered their existing (user-submitted) library of h264 video via HTML5. This whole brouhaha is about trying to convince them that they too, like openvideo.dailymotion.com and wikipedia, should use HTML5/Theora.
After all, most of their video content is user-submitted. They shouldn’t put their users at risk through using a royalty-encumebered format, especially, given the current h264 terms of use, for encoding.
Edited 2010-03-23 00:18 UTC
Then who pays for all the hosting? Who provides the infrastructure to make it work? BUSINESSES. The money has to come from SOMEWHERE. BUSINESSES like Google, Amazon, E-Bay, YouTube LLC… and if you don’t think Facebook and MySpace … or even better WordPress, Joomla and Blogger aren’t backed by companies and run as businesses behind the scenes…
… and 99% of the time if there’s no money in it, it’s going to fall flat on it’s face.
Its going to fall flat on its face if companies try to commandeer it and extract ongoing taxes for its use. Or at least, the companies’ use of it will fall flat on its face, if that is their model.
Ever heard of the dotcom bust? That was all about companies trying to use the web inappropriately. Their business model was wrong, and they all went bust. Meanwhile, the real internet itself just carried on.
The same will happen with web video if big business tries to commandeer the internet with DRM and pay-per-use proprietary codecs. The internet will “route around it”.
As for the relationship between FOSS and business (particularly software vendors) and your insane, illogical and frankly insulting attitudes to that, ponder on this:
http://www.mygnulinux.com/?p=470
Edited 2010-03-22 23:52 UTC
Those things are most certainly important to more people than you seem to mention. I’m no “die hard fight the power anti-corporate FLOSS zealot” (whatever that means), but I, too, am capable of looking beyond the corporate world (maybe because I’m not an American?), and see that between user and corporation, there’s a whole world too.
Hospitals. Charity institutions. Schools. Universities. And so on. You are claiming all those are “die hard fight the power anti-corporate FLOSS zealots”. Which is idiotic, to say the least.
How’d you like the vote this Sunday? Must sting, hmmm?
Yes. Taunt the screaming crazy man about politics. Because nothing bad could possibly happen.
Best way this situation can be resolved: Every browser supports both.
The main reason for supporting Ogg is that nobody should ever be forced to pay licensing to publish something on the Web. In this solution, nobody has to; you can use H.264 if you want (or need, like for phones, at least in the near future), but get ready to pay at some point.
All video currently in H.264 will just work and the video tag will be a success. People will realize that in 5 years licensing fees are going to hit them, and new video will trend towards Theora (especially as it improves). Even if Theora doesn’t get used much, knowing that there is a viable alternative will keep the MPEG-LA honest so they can’t pull a Unisys and start charging absurd amounts because there’s no alternative.
To implement H.264, browsers should just tie into the OS support. For XP users (and Linux users without GStream Ugly), provide an official plugin and suck up the licensing costs until XP dies (most Linux users install the Ugly plugins, so the cost of them is not really significant).
Basically Mozilla has to trade Microsoft, “we’ll support both if you do”. Opera will follow Mozilla, and developers need to berate Safari users and make them install the Ogg plugin for Quicktime until Apple gives in and supports Ogg out of the box.
Why are you assuming that H.264 encoders won’t improve during the next 5 years? The H.264 encoders have been improving over the past 5 years, so nothing should prevent them from improving in the coming 5 years.
At the end of the day, if the licensing fees are less than the costs of hosting Theora videos, companies will continue to use H.264.
Edited 2010-03-22 22:02 UTC
They probably will, but I’d guess that since there’s been more work put into improving H.264 encoders like x264 than has been put into Theora, there is more relative headroom for Theora to improve (just a logical guess, I could be wrong).
But yes, if the fees are small enough the savings of smaller videos and less bandwidth could outweigh them, but you have no long-term guarantees to what the fees are going to be. The MPEG-LA could decide to triple them the next time they renew licensing scheme.
But my (and Mozilla’s) point is about the “little guy”, meaning somebody with a blog or small businesses, where bandwidth/disk space are trivial compared to what the MPEG-LA could make them pay.
Edited 2010-03-23 01:09 UTC
Any idea of the current costs for users who already have videos on their blogs? If you outsource your videos to sites like YouTube (and I doubt this model will disappear), such won’t be the concern for the average blogger.
The small business on the other hand: any business that wants to get into streaming video now has to seek venture capital funding, even without the H.264 licensing fees. If you are streaming on your intranet, where bandwidth is ‘free’, there would be no point in paying for H.264.
“it seems like on the browser side of things, Theora is winning out; with a Theora video you can cover Firefox, Chrome, and Opera users – whereas with H264 you miss out on Firefox and Opera users.”
Yeah, because firefox + chrome + opera so clearly covers more ground than ie + safari + chrome.
IE doesn’t support HTML5.
Given that fact, your statement boils down to this: “because firefox + chrome + opera so clearly covers more ground than safari + chrome”.
True. That is clear. Very clear indeed, and perfectly correct.
from TFA:
“Now that Internet Explorer 9 has been let out its cage, we all know a great deal more about Microsoft’s position towards the video codec situation with the HTML5 video tag. Microsoft has chosen for H264”
However, the original statement was this:
“it seems like on the browser side of things, Theora is winning out; with a Theora video you can cover Firefox, Chrome, and Opera users – whereas with H264 you miss out on Firefox and Opera users.”
That is a present-tense statement.
In the present tense, IE (namely IE6, IE7 or IE8) does not support HTML5.
My statement, and that of the OP, are therefore both correct.
By the time IE9 eventually comes out, the situation may have resolved. If by then IE9 still didn’t support Theora/HTML5, it may be seen as deficient for that lack, just as Safari is deficient right now.
Does it really matter what I type since the only thing which matter is that you can claim theora is so great and leading even though it’s total bullshit?
Whatever, keep dreaming. And do your drugs.
Mere assertion and ad hominem attack on your part.
Meanwhile, here are some comparison files showing still frames.
The first set shows an almost-identical-quality video frame comparing YouTube’s h264, Theora Thusnelda and Theora Ptalarbvorm:
Youtube @ 499kbit:
http://people.xiph.org/~greg/video/ytcompare/bbb_youtube_h264_499kb…
Theora Thusnelda (current version) @486kbit
http://people.xiph.org/~greg/video/ytcompare/bbb_theora_486kbit_366…
Theora (Ptalarbvorm – new experimental version) @376kbit:
http://people.xiph.org/~greg/video/ytcompare/bbb_theora-ptalarbvorm…
I have got solid, objective evidence. You have got nothing but unsupported assertions, unwarranted insults, sour grapes and FUD.
BTW: here is more evidence of the improvements coming soon to Theora via the now-still-experimental Ptalarbvorm:
http://people.xiph.org/~greg/video/ptalarbvorm/
The parkrun video file at a slightly small filesize is significantly improved in the version made by Ptalarbvorm. (The improvement can best be seen in the trees in the far background). Meanwhile, the Thusnelda version is already almost as good as h264 at that same filesize.
Edited 2010-03-23 10:19 UTC
I don’t remember longer but the less polite version of my previous reply which I avoided would had been something like:
Just f–k off!
Point being IE 9 has chosen H.264.
Now go jack off with the kids in the sand box.
Edit: And it was never about the quality of H.264 vs Theora in this thread f–ktard. It was regarding whatever Mozilla should stick to their choice because IE9 would use H.264 or not, there I read the later as the more obvious news story (so rather: “IE9 specs are out: HTML5 video & H.264” or something such rather than his personal rant.)
I give no shit about your low-quality benchmarks. I complained on the line which made it sound like Theora had a wider browser support, which it definitely will not if current positions stand and IE decides to go H.264.
Edited 2010-03-23 10:22 UTC
f-off yourself, peabrain.
Point being: if IE9 eventually comes out with only h264 support, and cannot play Theora, then it will STILL be behind the times and deficient compared to contemporary browsers that can already support HTML5/Theora.
Vapourware, sonny Jim. IE9 does not exist. Theora does have wider browser support.
When IE9 eventually does materialise, if it cannot support HTML5/Theora, it will be deficient. Incapable. Behind the times. Poor. Just like IE6, IE7 and IE8 are now.
Edited 2010-03-23 10:28 UTC
Not that I want to get in the middle of your slugfest, but HTML5 has not been standardized, and won’t be for some time, so it has as much ‘vapour’ as IE9.
Regarding H.264, most video sharing sites support it via Flash, which has a much higher penetration that most browser plugins, so native support via a video tag is a moot argument.
Well, regardless of it being recommended as a web standard or not (it is being held up mostly by proprietary interests who want you to pay for h264), nevertheless wikimedia and openvideo.dailymotion.com work perfectly well right now with HTML5/Theora.
http://openvideo.dailymotion.com/en
There are over 300,000 Theora video clips on openvideo.dailymotion.com.
http://commons.wikimedia.org/wiki/Main_Page
http://videoonwikipedia.org/
http://openvideoalliance.org/2010/03/lets-get-video-on-wikipedia/?l…
There is a library available called the Kaltura HTML5 media library.
http://www.kaltura.org/project/HTML5_Video_Media_JavaScript_Library
My bold.
You can get the Miro Video Converter for Windows or Mac (Linux already has plenty of support for Theora):
http://www.mirovideoconverter.com/
This is anything but vapourware. 300,000 Theora video clips clearly aren’t vapourware. You can get all this better Theora video for the web performing via HTML5 right now, today, on any platform, for any browser, at zero cost.
Clearly not vapourware.
Flash is not a standard, either.
Flash does have good support, but the very problem is that it uses a proprietary h264 codec. That is not a feature, it is a liability. Using it carries a risk of being sued, and being made to pay ongoing license fees … even eventually for ordinary individual end users.
Now is the time, and the perfect opportunity, to drop h264 and Flash on the web.
Edited 2010-03-23 12:43 UTC
You’re dancing around a legitimate objection. As content hosts/providers like YouTube start trying to decide what format they’ll ship video in, it’s entirely likely that they’ll give more weight to IE9’s H264 playback capability tomorrow than to Firefox’s Theora playback today — especially given that there’s no logical reason to think that Microsoft won’t ship an H264-playing IE9 in a reasonable amount of time. And if they settle on H264, then it’ll be FireFox that looks “deficient and behind-the-times” when it can’t play youtube videos (or gets the Flash fallback).
I’m not saying any of this is a good thing — I’m not happy about it. But it’s also pretty much the way things are.
YouTube are experimenting at this time with HTML5. Currently, in that experiment, they use video encoded with h.264 codec, becuase that is what their current video “stock” is. This does not mean that Google are going to go with h.264 in the final analysis. If you have the compute power that Google has available to it, it is relatively easy to convert to another codec format.
So apart from the fact that Google has a vast array of h.264-encoded video which was convenient for its trial of HTML5, the only other objection that Google had to Theora (at the time of their initial HTML5 experiments) was that in their view Theora would take too much bandwidth. That view, however, was stated when it was true, when Theora was at version 1.0.
The original objection that Google had to Theora no longer applies. Theora at version 1.1 (Thusnelda) is currently competitive with h.264 in terms of its quality/bandwidth trade-off. (Experimental versions of Theora are significantly better, but at this time they have to be counted as vapourware, because they are not released yet).
So, h.264 will cost Google YouTube in terms of licensing, and lack of control (their business has dependency on the whims of MPEG LA). Theora won’t have either of those problems. Either one costs the same in terms of bandwidth. Theora has broader support in browsers right now, and it is trivial to add support for it to other browsers (since no licensing constraints apply). The only cost (which is a cost for Google) is in converting the video format for their existing corpus, which requires only compute power, which Google has in spades.
You do the math.
Edited 2010-03-23 22:38 UTC
I’m doing the math. If you go ahead and count expected IE9 deployments for h264, I’m seeing much wider-spread support for h264 than Theora. I’m seeing more of Google’s customers withing reach using h264. And as deathshad0w pointed out — galling tho it is to agree with him on any point — a licensing cost of “pennies on the dollar” isn’t nearly enough of a cost to compel Google to go with Theora, at the expense of loosing every IE9 user.
I’d love to see an open format get settled-upon as much as you would. But it’s looking less now.
Two points:
(1) Theora has no constraints on its distribution. Anyone can implement a player (even Google/YouTube themselves, just as Mozilla have done). Hypothetically, if an IE user were to visit YouTube after YouTube had changed to HTML5/Theora, then YouTube could simply give that user a link to a plugin, just as YouTube does now for Flash. The only difference is, because Theora has no constraints on its distribution, YouTube itself could host the HTML5/Theora plugin for IE, instead of relying on a third party such as Adobe. It need not even be a link, it could be a simple pop-up (on IE browsers) asking permission for the required plugin to be auto downloaded and installed.
No IE users need be lost at all. Because of the infamous deficiencies of their browser, IE users are already well used to having to install plugins.
(2) In pretending that the cost of h.264 to YouTube is merely “pennies in the dollar”, both deathshad0w and yourself are ignoring a critical point: control. If Google/YouTube decide to make themselves dependent on h.264 being on user’s browser, then not only do they create a potential serious ongoing liability for their users, but they give up control of their own business over to these parties:
http://www.mpegla.com/main/programs/AVC/Pages/Licensors.aspx
That is not exactly a list of “friends of Google”.
These two points are critical. Please do not ignore them.
Edited 2010-03-24 02:13 UTC
I think this would only apply to Linux. Seeing that the most popular OSes in the coming years (Win7/OSX) will have native H.264 support, browser vendors could opt to use the operating system’s native decoder.
Who controlled H.263 or VP6? For the past 5 years, many video sites were using them, and we didn’t care who controlled it or know what licensing fees were paid. I’m sure MPEG-LA and On2 got paid, unless we are to assume that they gave them away for free.
Which website didn’t require venture capital funding to stream online video? Streaming video is expensive, so one has to assume that licensing fees were part of the costs. If, as claimed, YouTube is spending over $300M/yr on their bandwidth bill, what is an additional $10M/yr for an H.264 license?
One point that seems to be ignored is that Google bought On2. Therefore, we can assume that they don’t plan on using Theora, but plan on using VP8 for their streaming video in the future. They have all the control, but not the ubiquity [many internet accessible devices can handle H.264].
Why would you think that? There are NO restrictions on distribution of Theora. None. Including distributing it to Windows users.
http://www.theora.org/
http://www.theora.org/faq/#14
My bold. Even commercial developers can write a Theora implementation with no restriction of any kind.
Mozilla distribute open-source Firefox to Windows users, and that includes a Theora player embedded within it, so just by having Firefox Windows users already have Theora.
There is absolutely nothing stopping Google from changing to Theora/HTML5, and offering any Windows users a plugin for IE that can play Theora/HTML5 in IE. Nothing. It could be done tomorrow.
Useless if it costs websites control over their product in trying to serve H.264 video.
Websites can offer Theora video as much as they want, no royalties apply, and they aren’t beholden to anyone for permission to serve videos.
There was no viable open choice then. There is now.
For a site like YouTube, it isn’t so much the cost, but the CONTROL that is the primary issue. YouTube cannot afford to have a hostile entity such as MPEG LA have control over YouTube’s permission to stream online video. That would be madness. Now that YouTube have a choice, MPEG LA lose control over YouTube. That is almost a must-do for YouTube.
Google bought On2 when Theora was not competitive. Theora is competitive now.
If Google release VP8 under similar terms as On2 released VP3 for development by Xiph.org, only then VP8 might become attractive. Right now, however, Theora is all but equal to H.264, and soon Theora (ptalarbvorm) will be significantly better than H.264. compared to that, VP8 is nowhere right now.
http://people.xiph.org/~greg/video/ptalarbvorm/
Edited 2010-03-24 11:08 UTC
I was thinking along the lines that: Linux can’t distribute H.264 decoders; OSX and Windows 7 are distributed with H.264 decoders. Windows and OSX are going to make up 95+% of the desktop market in the coming years. Are you going to invest in trying to install Theora decoders on those systems, or simply use the native H.264 decoder?
You do know that ptalarbvorm is vapourware?
With projections of smartphone adoption increasing and most (if not all) have H.264 chips, are you going to neglect that segment of the market? The trend is for H.264 adoption to increase, with the added benefit that it isn’t limited to the PC market. Neglecting H.264 would be akin to making a portable music player and neglecting MP3 because you didn’t want to pay the MP3 license. How many AAC/Vorbis-only players would sell?
Because a format is open doesn’t necessarily mean it will make you the most money. Everyone is waiting for Theora to become a credible alternative, whilst H.264 is working today. YouTube made a business decision based on what was the best available format at the time. Should they recode all of their existing videos everytime there is a new beta of Theora? Are we to assume that H.264 encoders have reached their zenith, and are waiting for Theora to surpass them?
Very well said.
If you are running IE, then in order to check out HTML5/Theora you will need to install support for it. Something like this will do the trick for you:
http://www.kaltura.org/project/HTML5_Video_Media_JavaScript_Library
If you have Firefox or Google Chrome, you don’t need this, as they already support HTML5/Theora.
OK, once you have got that going, check out some demos:
http://www.html5video.org/
http://www.html5video.org/demos/
(all kinds of filters & effects that Flash cannot do).
http://people.mozilla.com/~prouget/demos/mashup/video.xhtml
http://people.mozilla.com/~prouget/demos/DynamicContentInjection/pl…
As for quality, here is a demo of 720p
http://jilion.com/sublime/video
Nice. Quality. Sharp 720p resolution, still at a modest bitrate.
Enjoy.
Edited 2010-03-23 13:11 UTC
http://www.streaminglearningcenter.com/articles/ogg-vs-h264—round…
I can only imagine what your perverted vocabulary consists of in youtube.
“Oh kids in embedded H.264 video make me so horny!”?