[whatwg] Chipset support is a good argument

[Lino Mastrodomenico]

2009/7/6 Maciej Stachowiak <mjs at apple.com>:
> Here's an example of some markup that will work on a wide range of browsers,
> if you provide Ogg and MP4 versions of your video:
> <http://camendesign.com/code/video_for_everybody>. The MP4 version can be
> played either through <video> in browsers that support that, or by Flash or
> QuickTime or Windows Media Player. This actually results in video that works
> better in more browsers than Flash alone.

My first goal is to get as much as possible of my users to watch my
videos, but my secondary goal is to do so in a way that encourages a
sustainable video ecosystem on the web. I don't think that including
an easy and always available fallback to H.264 gives Apple and
Microsoft any incentive to support Theora or other free alternatives.

Moreover, while MP3, MPEG-4 ASP and H.264 are entirely fine and widely
used for pirated music and movies (which is illegal anyway), I prefer
not to put any H.264-encoded file on my website, things like these
scare me:

<http://www.streaminglearningcenter.com/articles/46/1/H264-Royalties-what-you-need-to-know/Page1.html>

and:

<http://www.streamingmedia.com/article.asp?id=11011&page=1&c=7>

Apparently the MPEG LA has not yet decided if and how much they want
me to pay from january 2011.

I'm sorry if this sounds like FUD, but I will not use H.264 until is
available without fees for the websites (guaranteed forever, not in a
we-may-change-our-mind-anytime way) and with an open-source-like
patent licence for encoding and decoding.

I'm not asking to the MPEG LA to kill their business, they can sell a
single patent licence to the open source/free software community for a
ridiculously high price. Insanely high. If it's applicable to derived
works, the community may well pay for it, and if it uses (for patents)
terms similar to the ones that the GNU GPL uses (for copyright) then
the MPEG LA will still be able to make more money by selling
"traditional" lower-priced licences to proprietary software vendors.

I hope they change their mind soon, but until the MPEG LA keeps H.264
"captive", it's simply not an option for me as a web developer.

> (Personally I'd recommend putting the H.264 <source> first instead of last,
> so browsers that support both H.264 and Theora will pick the higher-quality
> video.)

I use a high enough bitrate (roughly 0.2-0.25 bits per pixel, but it
changes a lot depending on the material and the resolution) that most
people can't tell the difference even on side-by-side comparisons. And
if they did I don't want to penalize browsers that chose to support
only Theora, since they IMO did the right thing for a sustainable
future of the video on the web.

Anyway I tried deinstalling XiphQT on my Mac and Safari doesn't play
the YouTube fallback inside <video>, so I'll include a small JS that
detects the situation and completely removes <video> leaving only the
YT <object> (BTW, canPlayType in Safari 4.0 seems buggy: it always
returns "no", even with XiphQT installed).

-- 
Lino Mastrodomenico