Safari supports everything that Quicktime does. You can add the support yourself, just like you could with Theora: http://xiph.org/quicktime/
http://perian.org is a Quicktime Component that wraps around libavcodec (ffmpeg) that has a pretty wide install base on OS X and checks for automatic updates by default every time it's loaded in an application. Google has a set of patches for ffmpeg on their site that will surely get picked up and automatically distributed before too long.
If Apple ever ships support, it'll be with Quicktime, not the braindead approach of linking it directly into the browser. IE9 has this approach too, just with a codec whitelist. Google Chrome bundles a pared-down libavcodec that you can swap out with your own. Opera does the same but with GStreamer.
In reality, the new version of Flash with VP8 support will get %95 penetration in a couple months after it's released, and fallback to Flash will continue to be how the <video> game is played. I wonder if Google's stopped charging Adobe for VP6 as a quid-pro-quo? It seems like that (and VP7 for Skype) were On2's sole sources of revenue.
Just curious: do you know why IE9 uses a codec whitelist? Why not just support any codec installed on the PC? Does Safari's QuickTime usage have such a whitelist?
Because the codecs are unsandboxed native code running in the browser process with full access to your user data. That you downloaded from the internet. To enjoy pornography and pirated media (and pirated pornography!). It'd be a perfect exploit delivery vehicle, and on a newly-prominent attack surface (see the recent @font-face exploits)
It's a much bigger problem on Windows because shitty DirectShow codec packs have a huge install base -- I wouldn't be surprised at a hundred million! It'll be a little better now since XP users won't get it, and the codec pack situation has cleaned up a lot in the last few years.
The IE team is doing the right thing for the situation they have. Anybody that cares is going to be using a different browser, their support for <video> just means that we won't have to use a Flash fallback for Vista/W7 users that allow updates.
Google is unlikely to prompt IE users to install the WebM codec on Youtube, since it'll work better with h.264 anyway.
Apple doesn't use a whitelist now, but they might in the future. Their attack surface is more constrained, since it's only the code they ship directly (the library, their default codecs, and the pro stuff they ship with Final Cut), Perian/libavcodec, and Xiph/liboggplay that are actually installed anywhere. No cambrian explosion of repackaged hacked codecs to deal with.
I wonder if a side effect of WebM will be to push Ogg Vorbis into wider adoption? Platforms that want to play WebM will also decode Ogg, so why not support the audio-only format?
Much like Theora, the primary reason to oppose Vorbis is political. It's not covered by patents and doesn't support DRM, so to content providers it's a significantly worse codec than MP3 or AAC.
If Apple or Microsoft were going to support Vorbis, they'd have done so years ago.
Not to detract from your general point, but Ogg and Vorbis support DRM to the same degree that any other container and codec do, it's generally an orthogonal concern.
It's main problem is network effects, in cases where network effects don't apply, e.g. game audio and web streaming audio, it has longe been used by big names such as Microsoft and Spotify.
There is now no reason for it not to be the standard for web audio, since you can even fall back to Vorbis in Flash 8 via a non-native decoder. It's also likely to get a boost in the pirate video market, even when paired with H.264 as a result of WebM introduction. It really could become mainstream for a variety of uses.
The biggest drawback in the last few years has been the lack of support in iPods. It will be interesting to see if Apple chooses to prevent you playing Vorbis audio on your iPod Touch even if it caves and allows it for web audio.
Vorbis has been around for over a decade now, and incorporated into numerous products with no successful patent threats yet. Xiph.org conducted a patent search and believes Vorbis does not infringe on any patents. And since US patents expire in 17 to 20 years, there is a shrinking body of patents that could be used against Vorbis.
No, apparently he does mean MP3, but don't ask me how a patent filed in 1997 covers a standard that was finalized in 1991. (Was there a delay between date of invention and date of application, maybe involving provisional patents? I do not know the details of these things.)
Do any of the browsers current support streaming with the HTML5 video tag?
So that you don't have to download the entire video before playback can start. And so that you can skip ahead and start playing something towards the end of the video without downloading the whole video.
So that you don't have to download the entire video before playback can start.
That's called progressive downloading, and it would be totally stupid not to do it.
And so that you can skip ahead and start playing something towards the end of the video without downloading the whole video.
That's called seeking, which is harder. Firefox can do seeking if your Web server supports byte ranges. It looks like Chrome does the same thing but I'm not entirely sure.
