Tue 11 Dec 2007
Open formats are the only way to guarantee an open web
Posted by J5 under Linux , Politics , Standards , News , Open FormatsIt seems Apple is not the entity fighting for a DRM-less world we thought they were and Nokia isn’t such a great patron for open source software either. For those who haven’t heard the news, lobbying by both companies have caused the W3C to strike a provision in a HTML5 draft which would make OGG Vorbis and Theora a standard media format on the web. What baffles me about this is having Vorbis and Theora media formats does not stop Apple or Nokia from distributing content using whatever proprietary format they wish. This just hobbles the web which had exploded thanks to open standards, open formats and open source. In recent years though the web has become more and more proprietary. Allowing companies to remove open formats from an open standard without nominating another equally open format to take its place allows special interests to control the development of the web. This effectively locks out a large portion of the small independent content providers from distributing next generation content such as video while moving us toward a world where a few entities control the distribution of all content.
There is a light at the end of the tunnel though. It is time for us to make the OGG family of codecs ubiquitous with or without the mandate of a standard. A large step in that direction is Firefox developing native OGG support. Easily installed IE plugins would go a long way also. It is also up to the community to post their content in an open format. The next time linux.com post a link to mp3 content that they produced and not have a Vorbis link associated with it I hope people take them to task. The same goes for people posting content on the various planets. Open formats are essential to a free and open web.
[read this post in: ar de es fr it ja ko pt ru zh-CN ]
December 11th, 2007 at 7:08 pm
[…] Read the rest of this great post here […]
December 11th, 2007 at 7:34 pm
Totally agree, good to hear this from a GNOME board member!
December 11th, 2007 at 7:41 pm
Nobody ever thought Apple was actually fighting against DRM? Sure hope not, because that would be beyond delusional.
Good post otherwise, agree totally.
December 11th, 2007 at 7:51 pm
I agree that having open standards/formats is a good thing. However, when these formats don’t measure up against the opposing technologies, there is no reason to use them.
Ogg Vorbis is so-so compared to the leading 3 audio formats, but Theora sucks balls. There is currently no encoding or decoding implementation that it’s really good, there are artifacts on playback on any player when moving the cursor in the timeline. And by their own admission, the developers used LESS efficient algorithms to go around patents.
Unless Theora becomes as good as h.264 one way or another, there is no point supporting Theora IMO. Web users want things that “just work”, not random artifacts and weird problems or inefficiency.
December 11th, 2007 at 8:14 pm
http://lists.w3.org/Archives/Public/public-html/2007Nov/0153.html
http://www.illiminable.com/ogg/
December 11th, 2007 at 8:16 pm
Eugenia,
While there is some truth to your assertions you are completely wrong about Vorbis. Vorbis has consistently measured up to all the other codecs out there and there is work to make the encoder even better. As far as Theora goes work is needed in decoding performance and encoding quality but it is far from bad (as can be seen by the the quality of Theora files at Red Hat magazine). The point is there is no other free formats out there that can compete with Vorbis and Theora. Having these formats supported does not force other formats from being used.
December 11th, 2007 at 8:30 pm
I agree with you that there is no other free format that can compete with OGG Vorbis/Theora in terms of user “delivery” formats.
But HTML5 is going to reach both FOSS enthusiasts as well as people who don’t give a rat’s ass about this. And these people outnumber the FOSS software enthusiasts.
Which is why it’s important to compare Theora to h.264 and acknowledge that it’s the not the optimal solution as a video format. It’s not a bad format, just not the best of the ballpark.
December 11th, 2007 at 8:37 pm
The spec never required Vorbis or Theora support. It merely suggested it. The wording was SHOULD, not MUST, which is important.
In the end, that standard would have been useless anyways. Even if it was there, Safari on Mac and IE on Windows would likely have no bothered to support the format, and it wouldn’t matter.
Out here in the real world, W3 standards are meaningless. All that matters is the subset of standards, the proprietary additions, and the bugs that browsers offer. The spec could praise Theora all it wanted, but if most professional video shops keep pushing WMV and QuickTime and such, or if even one of the three main browsers didn’t support, it wouldn’t matter a whit what the spec said.
December 11th, 2007 at 8:40 pm
Thanks for echoing our concerns, John!
December 11th, 2007 at 9:20 pm
Eugenia,
I guess my priorities are different than yours. I actually care about keeping the web free and open. It is the reason why the web was able to innovate and flourish to deliver what those people as you so eloquently put it “don’t give a rats ass” want.
December 11th, 2007 at 10:56 pm
J5: I worked on video encoding for long (way before I did GNOME or anything else), and I can only say that Eugenia is right, Theora really sucks balls. It doesn’t compare anything against 10-year old encoding formats such as MPEG-1, it doesn’t stand a chance even against something antique like MPEG-2. As for comparing it against MPEG-4r10 / h.264, I can just smile and shake my head.
Dirac may be a good free alternative to high-quality encoders, and I think other people are starting to see, e.g. all GStreamer work on Theora was abandoned to focus on Dirac instead, and it is obvious why. If HTML5 is to be a standard, it should promote high-quality and bleeding-edge technology, else nobody will adopt it, and people will use their own codecs instead anyway. Therefore, requiring Theora was wrong (and requiring Vorbis and Theora includes Theora, and is thus wrong). Requiring Vorbis alone would have been silly.
Unless there is a bitstream-stable codec with universal (Mac, Win32, Linux) implementations available and installed by default in computers from large suppliers, high-quality and good compression ratio, people will not use it, because better alternatives will exist. Let’s hope Dirac is that. If it isn’t, we’d better as hell get our ass moving if we want to play any role.
December 11th, 2007 at 10:56 pm
J5,
Eugenia is actually correct, here. I applaud you and your desire for a free web experience, but you are going about it all wrong. The way to get a free web is by convincing CONTENT PROVIDERS, not trying to get browsers to support the codecs.
Now, granted, having the specification list ogg/theora as a possible video format in this particular case would be good, because without it, there’s no hope.
BUT
There was really no hope for it in the first place if it isn’t up to snuff (which as far as I understand, theora simply isn’t). Plus it doesn’t support DRM which leads back to the need to convince content providers.
To content providers, DRM is a high priority item and until you can convince them that DRM is bad (hah!), you are up out of luck.
So, failing that, you need a codec that supports DRM in order to even have a chance in hell of convincing them… do we have a format that supports DRM? No.
Now, as to your complaint about Nokia & Apple not /also/ supporting the ogg codecs - they actually have valid concerns. It’s impossible to tell for certain that no one holds a patent that theora/vorbis /may/ infringe upon. Also recognize that the media space is HUGE as far as money is concerned, and that means that they are at a much higher risk of patent attacks because so much more is at stake.
It doesn’t matter if theora/vorbis/whatever doesn;t /actually/ infringe on patent XYZ (or that patent XYZ is bogus), it’s /still/ going to cost Nokia/Apple/etc a huge amount of money protecting themselves.
You work at Red Hat - you guys are being sued by Acacia (sp?) for supposedly infringing on some bogus patent, but it’s going to cost you guys millions defending yourselves even though it is bogus.
That’s just the way the world works, burying your head in the sand and pretending it’s as simple as adding support for another format at no cost to Nokia/Apple/etc is naive.
December 11th, 2007 at 11:07 pm
You know, it is funny. I was courting the same subject with a friend of mine (whom happens to be a Mac-user) over the OLPC XO and power consumption. His argument being that a ‘child in Africa would not care that the CPU is low on power consumption.’ My counter point being is that if the power consumption question (That being, how much can the XO hold and how long would the battery last on one full charge based on the CPU’s power requirements) was never asked, there wouldn’t be much that differs the XO from every other laptop on the open market. Which would clearly be unsuitable for the environment that these kids would be working in.
The relation here is that, regardless of quality, OGG Vorbis/Theora would insure that the author would own the content, or have some ownership granted to them. For the open web, as it stands… this is necessary thinking. The web at large is a canvas, allowing one to create many things (for good or for ill) and that users drive that creation. As far as I see it, OGG would give users the power to create and portray it over the web as they see fit. I can’t say that any other format in the audio/video world would allow for that. It something that might miss the masses complete, but if YouTube has proven anything about user-driven content… is something that all citizens of the net indeed applicate.
December 11th, 2007 at 11:44 pm
What a lot of people seem to miss is that this isn’t about just vorbis. Fine reject it but I get to call you out on it unless you actively work on an open format that I will be able to use. Here is the point, these big companies are not the only content provider. I am in fact a content provider and so are many other users.
Companies can do what they want with their own software but their move on a standards draft to remove open media formats allows them to use the same excuses for any standard that doesn’t mesh with their business plans regardless if the argument is legitimate or even their real motive. By the same arguments Nokia wouldn’t touch Open Source software because a) they didn’t buy a license from someone b) for the most part they didn’t write it themselves. But hey they like getting work for free (not to say they haven’t contributed good Open Source work but it kinda waters down the patent argument).
I want to be able to distribute content unencumbered. Companies can use whatever they want but at the end of the day I would like a defacto open format that anyone can view. Looking at the video at Red Hat magazine quality is actually quite good. If it isn’t going to happen in a standard I have already talked about the alternative such as Firefox getting native support (WebKit has it in a branch too). If Dirac proves to be better then we can move to that in due time.
It seems that I am not the one putting my head in the sand. I understand the issues involved, head in the sand would imply advocating just Theora and nothing else. However giving up, letting the companies have only their way without thinking about the small content providers, is putting your head in the sand. As I have said in my post having this as a recommendation in the standard does not restrict them from shipping their own formats.
December 12th, 2007 at 12:01 am
Except that you still have your head buried in the sand wrt patents.
December 12th, 2007 at 12:43 am
J5: you still didn’t address the quality issue. No sane content provider, regardless of how small, will prefer a crap format over something equally crappy but much simpler a la Youtube.
Theora is crap, that needs fixing.
December 12th, 2007 at 1:02 am
“I am in fact a content provider and so are many other users.”
I fail to see how the suggestion to use Theora in the HTML5 spec really matters to you as a content providor. If you want to provide content in Ogg Theora, go ahead. I can almost guarantee you that the element will not _just_ support whatever formats are in the spec; your gstreamer machine will support any codec you have plugins for in a video element, and the same will be true for Apple with QuickTime plugins and Microsoft with WMP plugins.
The point is, it just flat out does not matter what the spec recommends for video codecs. The spec probably shouldn’t recommend anything at all.
What you really want is for content providors to produce non-DRMd content and for OS distributors to ship Free codecs with their media stacks. A small optional format recommendation in HTML5 will do neither of those things.
December 12th, 2007 at 2:17 am
So, I did my own tests about the quality of the two, here’s the result:
http://osnews.com/story.php/19019/Theora-vs-h.264/
December 12th, 2007 at 3:54 am
J5,
Agree with article and comments 14 completely.
Eugenia,
Can you find some more time to download/install theora 1.0 beta 2 and do the tests again.
And since I am not sure how good is ffmpeg2theora can you try ripping some video from a DVD using like Thoggen.
The reason I am asking this is that I am using theora 1.0 beta 2 for over a month for ripping my DVDs and I am happy with the quality. If you also think same as me then probably we should give theora more time for improvement in quality and performance. Hopefully it will equate (if not surpass) h.264 in near future.
December 12th, 2007 at 4:20 am
No, what you say does not make sense. I used an HDV stream which is MUCH better quality than a DVD, and I even offer the original video source file to try it if you want.
As for ffmpeg2theora 0.20 it was only released a few days, so it probably has Theora 1.x in it anyway.
All you have to do is try it yourself. You have the source video, you have the command lines used, and you have an encoder, so you need nothing else to prove me wrong.
December 12th, 2007 at 4:29 am
ffmpeg2theora just uses theora apis. It does not include theora library in itself. So you still need to install theora 1.0 beta 2.
Apart from that I will try to re-encode the videos on my machine sometime this week and post the reply on your blog.
December 12th, 2007 at 4:49 am
If it didn’t link against the theora library it wouldn’t work.
December 12th, 2007 at 6:08 am
[…] Check it out! While looking through the blogosphere we stumbled on an interesting post today.Here’s a quick excerpt It seems Apple is not the entity fighting for a DRM-less world we thought they were and Nokia isn’t such a great patron for open source software either. For those who haven’t heard the news, lobbying by both companies have caused the W3C to strike a provision in a HTML5 draft which would make OGG Vorbis and Theora a standard media format on the web. What baffles me about this is having Vorbis and Theora media formats does not stop Apple or Nokia from distributing content using whatever propriet […]
December 12th, 2007 at 6:18 am
Ronald S. Bultje: Huh. No sane content provider? What about those crappy Flash movies all over the internet?
Most people DON’T care and people DON’T see any difference (and yes, I have tested Theora and DivX.
WHO cares is content providers. Some content providers want DRM, some don’t. That’s fine. But providing one fall-back format which is royality free and uncontrolled. For now, Theora is “good enough” for that.
I understand all ramblings about how cool MPEG-4 or any other “new format” is, but in fact, this doesn’t matter, because they are proprietary and can’t be used by all so freely.
Anyway, J5, congrats about getting elected. Keep up doing your work, you rock.
December 12th, 2007 at 8:06 am
@Ronald: I do not think the claim of ‘all GStreamer work on Theora was abandoned to focus on Dirac instead’ is true or correct. The Theora plugin in GStreamer is quite mature at this point and thus the amount of work that needs to go into it is little. That said Theora has been getting RTP support in GStreamer and it was the first codec to get reverse playback support.
So while Dirac is definitely a contender for the future, Theora has not been made into abandonware in GStreamer.
@Eugenia: The success of Flash video was not caused by great quality. Their audio/video sync for instance was horrible for a long long time and the image quality with h263 which still most seem to use is not impressive.
Content creators will use Theora if they know that by making content available in Theora they will be able to reach a high percentage of web users out there, without needing an extra download. Cause that is the primary concern of content providers on the net currently, how to distribute content to as many people as possible without requiring an extra download. Because as soon as you demand an extra download you lose half your viewers.
December 12th, 2007 at 10:33 am
Putting a format where you have to pay license fees to use it in a standard is NOT the right thing. Theora might be crappy as hell. I don’t care. If there is no fee-free alternative it should go in the standard. The industrie might as well go ahead and create an really open format themselves and put it in.
But here is the point: You can’t have an open format with DRM. If the format is open who hinders me from working around the DRM? The result is as long as they want DRM in the standard they won’t allow an open format. Regardless of the quality.
December 12th, 2007 at 10:58 am
Alex, that is not correct. There is nothing hindering you from encapsulating an open format in DRM. And the legal situation around it would be no different than any other DRM’ed media.
December 12th, 2007 at 1:04 pm
Onkar: Ogg Theora is just On2’s VP3 codec, which is ancient and only comparable to MPEG-1’s quality. While that may be “good enough” for you, it is not good enough for the big content providers which want to provide HD content (and could easily argue “not good enough” for even pre-HD content).
December 12th, 2007 at 1:06 pm
> Content creators will use Theora if they know that > by making content available in Theora they will be > able to reach a high percentage of web users out
> there, without needing an extra download.
Except that this isn’t true… not many currently existing players (used outside of the Free SOftware community) actually support Ogg Theora.
Ogg Vorbis is far better supported out there and even it isn’t as widespread as it needs to be.
So this argument is bunk.
December 12th, 2007 at 4:26 pm
I’m frankly appalled by the negativity here.
I just wanted to echo J5 in saying that anyone savvy enough to be reading this blog should be publishing their video in Theora, and encouraging Gnome developers to do their screencasts etc. in the same way, embedding them as with HTML5 video tags.
Your users can already get Firefox and Opera builds with support, Safari nightlies work with the XiphQT codec plugin. The VLC plugin works for other situations and Java is a fallback if you use the mv_embed javascript:
http://metavid.ucsc.edu/blog/2007/06/07/html5-video-the-future-is-now/
If there are problems, whether it’s creating, editing, subtitling, streaming or whatever then we, the technically capable, need to be finding them and fixing them.
This isn’t about technology, it’s about free culture and there’s nothing other than Ogg Theora/Vorbis that comes close to fitting the bill.
Some more info on Theora progress:
http://web.mit.edu/xiphmont/Public/theora/demo.html
December 13th, 2007 at 1:51 am
Think about a long term approach. With the project like OLPC, using an uncumbered patented format/codecs is a good alternative. What some posters kept forgetting is Theora is another open format when videophiles and other developers are welcome to participate to the development.
One has complained about the use of outdated codecs On2’s VP3 but for the mass public, it is not very important. Let’s face it, having a high quality video rendering will require more CPU process. As the OLPC started to seed on some developing countries such as Uruguay, Peru, Brazil, having a decent quality of video will be enough.
The real problem is the maintain of status quo for several industries unwilling to get involved in a format like Theora. I think having an open format as mandatory wil greatly bring innovations. We need to think not only in the industrial view but also worldwide. So far Theora is the suitable solution.
December 13th, 2007 at 8:11 am
@Eugenia, Roland and all the rest praising video quality:
HD *sucks*. I mean, HD is simply a marketing buzzword, or for video-quality aficionados. The “dumb mass”, which we always refer to when arguing about anything that’s going to be widespreaded, just likes DVD quality. In many fields in technology we’ve reached the “it’s just enough” point, where a better quality doesn’t matter because it couldn’t be appreciated. There are millions of people listening to their crappy mp3 in their crappy player, and there are thousands of audiophiles screaming because of this. The same applies to video encoding, and you’re on the audiophile’s side.
December 13th, 2007 at 8:23 am
@Jeff: Dude we are having this the discussion in the context of built-in browser support for Ogg formats. If your browser support Ogg you don’t need a separate download.
December 13th, 2007 at 11:54 pm
“Ogg Theora is just On2’s VP3 codec, which is ancient and only comparable to MPEG-1’s quality. While that may be “good enough” for you, it is not good enough for the big content providers which want to provide HD content (and could easily argue “not good enough” for even pre-HD content).”
More incorrect statements than there are sentences!
“Ogg Theora is just On2’s VP3 codec”
No, it is built from it. The formats have diverged in some significant ways.
“and only comparable to MPEG-1’s quality.”
Bet two cases of good Lambic on that, name a neutral arena, and you’re on. Your weapon can be any MPEG1 encoder, mine will be our Theora encoder. Bring the beer with you, you won’t be leaving with it.
” “not good enough” for even pre-HD content).”
you.
tube.
FLV is the most dominant format on the web today. That hardly bodes badly for Theora.
But you’re right the old encoder is inefficient. That’s why we’ve built another.
See http://web.mit.edu/xiphmont/Public/theora/demo.html if you think I’m deluded and believe the current Theora can do no wrong.
December 20th, 2007 at 1:35 pm
“The next time linux.com post a link to mp3 content that they produced and not have a Vorbis link associated with it I hope people take them to task.”
Don’t worry, they do, usually rudely and loudly. And then there’s the deaf guy who demands transcripts of every podcast we do.
Message received. We must double our cost on every podcast to make tiny bits of our audience happy, so we must only do podcasts on hugely popular topics or simply not do them at all.