"If you read a lot of books, you're considered well-read. But if you watch a lot of TV, you're not considered well-viewed."
- Lily Tomlin -
        Syndicate

* Atom 1.0
* RSS 2.0

Videopia the Book for only $9.99! 236 pages, all original content on EVERY topic from light to copyright.

"In my opinion, the content is worth much more than you are charging for it..."
                            - David B.

"This booklet is a great resource, and worth many times the price. Heck, it's worth four bucks just for the entertaining read alone..."
                            - Shane G.

"This is the best explanation of
Internet video I have ever read."

                            - Mark H.

Theora Beats H.264 PDF Print E-mail
Written by D. Eric Franks   
Thursday, 07 May 2009 21:11

H.264 by any other name (e.g., AVC, MPEG-4, WMV, MP4, DivX, etc.) is THE dominant video encoding format, from mobile phone video to 4K cinematic distribution. It's the perfect format. Well, except that specific implementations are all owned by someone and there are always license fees associated with it, which we all pay when we buy encoding software or playback devices, although the cost is hidden in the price of the product. Fair enough: it's a spectacular format that is worth the price, in my opinion.

Granted, it'd be nice if there were something we could use that was as good and free. You know: like the open source PNG format, which combines all of the best characteristics of JPEG and TIFF and other "owned" formats. Just as good, in some ways better, and entirely free. Microsoft doesn't own it. Apple can't threaten us with it. Free.

The PNG of video is OGG Theora. Open source. Free. Quality? Meh. It's probably fine. But H.264 is SO good, that it makes you wonder: Who cares about Theora? Well, Christopher Montgomery at MIT does, apparently (although he works on the Xiph project, so not a big surprise there). A new PSNR study shows that Theora is BETTER than H.264! And they use clips from Star Trek to prove it! Holy crap: it's a video nerd perfect storm!

But what does this mean for content producers (a.k.a., me and you)? Well, to be honest: nothing. First, my analogy to PNG is a good one. I've been saving still image files to the PNG format for over a decade, primarily because it is lossless and the compression is good on it. Eventually, TIFF with LZW compression caught up to the open source PNG, but I still stick with PNG. Why? I dunno. I guess I feel good about using open source code, but, quite frankly, it's lossless and it's compatible so... why not?

OGG Theora, unfortunately, is not so compatible. In the early days of PNG, it had compatibility problems with browsers (at the time, Explorer and Netscape... remember Netscape?) as a distribution format. But I wasn't using it as a distro format, I was using PNG for archival purposes. Still do. Theoretically, PNG is superior to JPG for distribution, but I never, ever post PNG images on the Web. It may be extremely compatible today, but it's not AS compatible as JPG. OGG Theora, on the other hand, is almost completely incompatible with almost everyone's computer. Except super nerdy open source Star Tek fans.

It doesn't matter how good the format is or even how popular it becomes with consumers. There's only one thing that could happen to make Theora more compatible and that would be if it was picked up by software and (especially) hardware manufactuerers, perhaps in an effort to save licensing fees. You'd need a Tivo or a iPod or an NVIDIA to take the format and run with it. Then you'd need a Sony or an Adobe to support it in their editing apps. That's a steep hill to climb, but stranger things have happened, and it is possible that some company might come along and decide that it doesn't want to pay MPEG licensing fees with a proprietary distribution format ("proprietary" has rarely stopped Sony before).

Finally, it's important to note that PSNR analysis of video (which is what was used to measure the quality of the video) is an objective measure of fidelity and is not a subjective measure of whether a video actually looks better or not. This is important, because our mushy analog gray matter and eyeballs are not very objective. So PSNR is a good way to make quantitative measurements, but it doesn't necessarily say anything about the quality of the video as perceived by human beings. The authors of the analysis admit as much "It's important to note that PSNR is an objective measure that does not exactly represent perceived quality, and PSNR measurements have always been especially kind to Theora."

In any case, we can say that this development is "Fascinating" and it'll be interesting to see if the Xiph codecs can gain traction. At a minimum, these technologies and the fact that they are competitive with licensed codec keeps us safe from the big giant Evil Corporations: Without the threat of Theora, who knows how much we'd have to pay for H.264?

References:
* Theora: Thusnelda project update 20090507
* Blogging Theora - more awesome illustrations of DCT, motion detection and other details of exactly how the codec works, accessible to laynerds
* http://www.xiph.org/