Thursday, June 17, 2010

Should universities worry about H.264 video patent licensing?

The short answer appears to be: almost certainly no :)

This surprises me. For a long time I've been under the impression that universities encoding free-to-use but non-public content into H.264 would be liable to pay licence fees. Annoyingly small licence fees in most cases, but payable all the same. This doesn't appear to be a serious concern any more.

For those who don't know, H.264, a.k.a. AVC, is the de-facto standard "codec" for compressing video content these days. It is used by iPods, iPhones, YouTube, TV set-top boxes - you name it. It's ultra-scalable, handling mobile phone to near cinema quality HD video without being lossy, and can be encoded and decoded by a plethora of low cost chips. It's also part of the MPEG-4 family of standards, and so non-proprietary. This is in contrast to, say, the ON2 VP6 codec used by Flash video, which is still the best choice for distributing video playable to older computers, and is the default codec you are receiving when playing YouTube videos, for example. Non-proprietary does not mean free however, and use of H.264 is bound by patents and licensing agreements, collected together by the industry in the form of the MPEG LA AVC Licence Agreement. This agreement was updated this February. I've been looking at the new agreement again today and the picture is better than I thought. Here's the key section:
In the case of Internet broadcast (AVC video that is delivered via the Worldwide Internet to an end user for which the End User does not pay remuneration for the right to receive or view, i.e., neither title-by-title nor subscription), there will be no royalty during the first term of the License (ending December 31, 2010) and following term (ending December 31, 2015),
So, free H.264 internet content incurs no royalty, whether it is public or not, until 2015. Nice. After that? The agreement continues thus:
after which the royalty shall be no more than the economic equivalent of royalties payable during the same time for free television. ...  The first term of the License runs through 2010, but the License will be renewable for successive five-year periods. ... Royalty rates applicable to specific license grants or specific licensed products will not increase by more than ten percent (10%) at each renewal. 
So, it could jump to the same as free television rates - which I'll come back to - but even this can't increase by more than 10% at each renewal, so there's no fear of large price hikes in future which was previously a legitimate worry. If they did decide to price internet broadcasting the same way as free television, what's the hit? Here's where admittedly it does get a little more complicated.
where remuneration is from other sources, in the case of free television (television broadcasting which is sent by an over-the-air, satellite and/or cable Transmission, and which is not paid for by an End User), the licensee (broadcaster which is identified as providing free television AVC video) may pay (beginning January 1, 2006) according to one of two royalty options: (i) a one-time payment of $2,500 per AVC  transmission encoder (applies to each AVC encoder which is used by or on behalf of a Licensee in transmitting AVC video to the End User) or (ii) annual fee per Broadcast Market starting at $2,500 per calendar year per Broadcast Markets of at least 100,000 but no more than 499,999 television households, $5,000 per calendar year per Broadcast Market which includes at least 500,000 but no more than 999,999 television households, and $10,000 per calendar year per Broadcast Market which includes 1,000,000 or more television households.
It's a mouthful, but the above gives us the choice of two options, and the second is cheaper as it costs nothing unless you're challenged by someone seriously arguing that your broadcast market is over 100,000 users. If your university is getting those kinds of viewing figures out of your academic content - good luck to you! Even then the liability would only be $2,500 per calendar year, rising by only 10% every five years, which is hardly anything to lie awake at night about.

Yes - celebrations all round - H.264 appears to be free for universities to use. But should universities be archiving precious video content in H.264? This is really a question for library services like the National Archives, to which my university is affiliated, or JISC's digital media service. I have looked, but I have yet to see clear guidance on this. If anyone can provide a definitive answer I would certainly appreciate learning about it. My own view currently is that digital video content should be archived in it's "virgin" or "raw" format if possible, and H.264 at HD ("1020p" or equivalently 10Mbps) and broadband web-friendly (500kbps 700 kbps) bit rates. (Wikipedia puts these bit rates nicely in context.) My view on this may change if WebM takes off. This codec, which is comprised of VP8 video codec from ON2 (which Google bought), Ogg Vorbis audio, and the Matroska container format, is entirely royalty free. It's the new Ogg Theora, if you will. It promises to be almost as good as H.264 performance-wise and with Google's backing, and YouTube's by extension, it's sure to become the standard HTML5 video format of choice in years to come with native browser support. As such WebM should be the perfect choice for playable archive formats, but it's early days as WebM was only announced last month, and so encoders are still very hard to come by Even so, I believe that only the continued success of H.264 in the existing marketplace, and possibly Apple's anticompetitive tactics witholding WebM from iPhones, are likely to hold WebM back. Indeed the next few years could see a pitched battle between Apple the movie mogul promoting H.264, and Google the "we're not evil" hippies promoting WebM. It's yet another place where these titans are currently clashing.

So the bottom line is, universities can generate and distribute H.264 video content that is "free to view or receive" without incurring any royalties at all until 2015, and after then we could be liable to pay swallowable annual fees that may increase no more than with standard inflation, but this is very unlikely. The WebM alternative will probably provide an even better way of distributing totally free HD and web-friendly video in the future. Whichever of these wins out, the good news is that producing free to use video content for the web is only going to get easier.


Anthony Leonard said...

A quick follow up. According to the same agreement quoted above, if your university has H.264 encoded video content longer than 12 minutes that it wishes to sell "title-by-title" for more than $1, then it is liable to pay $0.02 per sale to MPEG-LA. That is an annoyingly small amount! The overheads would be in managing this process, rather than the cash involved itself, and here's where WebM makes sense again as an industry-wide friendly format that is totally free to use for any purpose, commercial or not.

Stevie G said...
This comment has been removed by the author.
Stephen Gray said...

Some great points well made Anthony. To answer your question on 'archival' formats (with my JISC Digital Media hat on), we'd advise uncompressed video (and in 4:4:4)be archived in an ideal world. That said, there are cases in which this isn't possible and H.264 might indeed become an archival format.
For instance, an archival policy which specified preservation of the 'originating format' might easily result in H.264 (either shot in-camera or acquired as a legacy item)being archived. Your 'virgin' format is a good rule of thumb for archival video. We've got downloadable samples on our site for anyone who wants to see different video formats in action.

Anthony Leonard said...

Thanks Stephen. Agree that any digital material should be archived in as close to its original state as possible, and that H.264 or something else is only useful for 'preserving' should support for the original format fall away, and/or republishing on the internet.

I've made some comments on the JISC Digital Media's Choosing a Digital Video File Type page regarding the need for clear guidance on what is a good choice of video file type and compression bit rate for internet publishing. I've said 700kpbs is best there based on recent local discussions, and so I may have to edit this post to reflect this slightly higher figure.

Thanks for being there by the way. JISC DIgital Media is an absolutely a fantastic resource for us.

Anonymous said...

simply stopping by to say hey

Anonymous said...

Wonderful web site. Lots of useful info here. I'm sending it to several pals ans additionally sharing in delicious. And obviously, thank you for your effort!

Look at my weblog: flexcare toothbrushes