So news is out that the W3C won't mandate what codecs must be supported by a browser as part of the HTML 5 spec.

They state:


Now the bit there about Apple worrying about patents in Theora peaks my interest, but only because I had an actual IP lawyer say the same thing to just a little while ago, but just that he has "concerns".. I'm starting to think lawyers every where say that about everything just because it leaves the door open for them to make a few extra dollars (hello Mono).

But that aside, I wanted to talk about another aspect of all of this:

One of the things that really bugs us Linux guys is how the fanboi tards of Apple and Microsoft (does MS have fanbois besides guys like Paul Thrott?) is the way they latch on to buzz words from he company, and then regurgitate them when ever they are challenged for details, or are preaching their path to salvation to someone when they really have no idea of the real technology behind that buzzword or if it truly is that great of a implementation.

But here's the thing everyone, I'm starting to really come to the overall conclusion that OGG is just becoming an OSS buzzword. Hear me out:

1. OGG is not even a codec. It is a container that can hold a codec... You could put MP3 audio inside an OGG container. But you hear people chant on about OGG and not Vorbis or Theora. When someone talks about Ogg being a "good codec," it demonstrates the kind of impractical, blind bias for free-sounding buzzword projects.

2. Keeping things focused on the video side, since that's what the <video> tag has the most potential for, lets discuss the issue of quality.

I look at video codecs in a cost per-bitrate kind of math. As in, if to make video X look decent at say 960x540 I need to have that video in 512kbps in one codec, and 768kbps in another, that codec that requires a higher bitrate is going to cost me more in bandwidth since it will have a large file size, transfer time for my self to upload and my viewer to download, and a higher cost in storage for my self, my server, and my viewer.

The original Theora codec came from On2, and was called VP3. On2 created VP3 in an attempt to compete with H.263 and is a very old design. To make matters worse, Theora has recently seen some really great improvements in it's bitrate to quality ratio but because these improvements are not official, and not yet shipping in any software encoder that support Theora (without building your own source) we don't get to see the improvements.

Even with these improvements, after real world tests by people who know how to get the best out of each type of codec Theora still falls flat. It simply is an old codec designed for last generations video.

See these examples comparing the latest Theora to the latest x.264 and VC-1: http://cid-bee3c9ac9541c85b.skydrive.live.com/browse.aspx/.Public/BBB|_Compare

3. H.264 video with AAC Audio is a great combination that is here today, and has a fantastic Open Source implementation in x.264.

This is the thing that really stumps me. x.264 is damn good. Amazing file size for quailty. I can put out a 720p (that's 1028x720) x.264 encoded video with a smaller file size than a Theora 540p (960x540) video. If you follow me there, I am saying I can put out an HD video and a standard deff version in Theora, and the Theora will have a larger file size than the HD version. On top of that, the Theora codec loses more color detail, and won't play back on a fraction of the systems, when you consider non-Linux desktop machines and devices.

Does h.264, AAC, or the OSS implementation of x.264 have patent issues? Maybe... Who cares? As a viewer of a video, you are not liable for any infringement, it's possible the producer of that content could be. But even as a producer, I'd have to be a crazy person with a tinfoil hat to think that some coalition that represents dozens of companies worth billions of dollars is going to come after little old Chris Fisher. It's just not worth it to them. Not to mention by the time they got down my driveway I'll have already apt-get remove'd my gstreamer-ugly codecs.

The other flip side of the up hill argument that OGG Vorbis and Theora fanbois have to fight is that professional content producers don't have to worry about the patent issues ether. When you buy Apple's $2,500 Final Cut studio, or Adobe's near $3k CS4 suite, they come covered. Apple and Adobe (and all the other makers of those types of software) have bought and paid for the distribution license.

From the angle of the hardware playback manufactures (like say, Popcorn hour, or TiVo), the cost for them to legally have a hardware based chip decoder is well under $1. So they look at the situation and thing, "well... That lawyer guy we pay $475 an hour says he has concerns... Or we pay less that the cost of a cup of coffee and we have a legal contract that says we are covered with h.264!".

I don't understand why we let these patent concerns in the OSS world cock block us so often, and then other times blatantly ignore them. Especially given our proven track record of being able to re-implement (ACPI, SysV, VFAT), adapt and move on so quickly.

You see it in every mainstream device running Linux today, The Palm Pre, the TiVo, the Android phones, on and on, all of these Linux devices support tons of codecs except for Theora. HELL - The new Popcorn C200 is DROPPING support for Theora from the previous model.

I feel like just shouting out on the top of a mountain MOVE ON everyone. Ogg Vorbis and Theora need to DIE. They just don't hold up to AAC and X.264. They are practically dead right now anyways. Let them go in peace!

ANDD.... </RANT>

Now allllll of that said, I still want to implement Theora video using <video> tag for browsers that support it, with a fallback to flash for those who don't. WHY after all of that above? Because I want make the content that we produce available in as many ways as those consuming that content want it. Plus I work my ass off to make the Ogg Theora versions, and they get a tiny tiny fraction of the downloads... So... I hope the <video> tag will make my efforts a little more worth while.

Let me know what you think... Am I way off base? Have I given you something to ponder? Or maybe you are way ahead of me, ether way let me know!

-Chris
(Cross-posted on my blog, bin-false.org)

Also commented on Chris's blog:

Great post! People talk about software patents being evil and such, and I agree that there a lot of patents out there that are. However, compression is a very academic endeavor requiring much serious research and development.. If there exist such a thing as a non-evil software patent, surely this would be a candidate. I have no problems with people using h.264, but I do love the fact that there are unencumbered formats available.

On a separate note, I would like to mention that there is another free video codec out there worth taking a good look at: Dirac.

YES, why is there not more buzz around Dirac, it seems like a codec built for today and the next few years.

-Chris

http://saintdevelopment.com/media/

"The development of Thusnelda was supported by Red Hat, Wikimedia and Mozilla."
Fedora: Features/Thusnelda

Monty's Journal
Theora: Thusnelda project update 20090507

I down like a clown for Thusnelda too, I just want us to pick one, and get behind it as a community. But the one we pick, as gotta be at least as good as x.264, otherwise, it's just a big fail.

-Chris

Personally I think that the media stuff should be a plug-in. It makes for a leaner browser, if you are only just doing biz apps you can use a more modest machine.

And while we're at it, let's get rid of <img> too! If you want a crippled web experience, there are plenty of light-weight browsers that don't support <video>, and even ones that don't support <img>.

I am not against video or other media. I think it should be available when you need or want it. Plug-ins or modules give you more freedom to choose. I do not want to be relegated just because my needs might be different. If you can choose what apps via synaptic or the like you are using, then why not be able to choose the plug-ins that meet your needs in a browser. That makes more logical sense to me. Thall shalt be modular....

It would certainly be entirely possible for a browser vendor to implement a solution for that. There's no reason why they can't.

As a content producer, you want to be able to put something out and be sure that as much of your potential audience can consume it. This is what the standards try to accomplish.

I think OGG is a victim of its own design. I just want to punch some one in the face every time I talk about OGG in general as if it's a codec and I get smugly corrected by being told "OGG is not a codec, it's a container", with the inevitable mp3 example. It makes me want to not use it.

I support the use of the best codec for the job. Computers are a tool to perform a task we need done. You use the best tool for the job, if that's not Theora then so be it.