Your (linuxeventually) acronyms are flumoxed.

Video: VP8 vs H264 vs Theora
Audio: AAC vs MP3 vs Vorbis
Container: OGG vs Matroska vs MP4 vs 3GPP vs AVI

WebM is VP8+Vorbis+Matroska.

HTML5 OGG doesn't make a whole lot of sense. Both OGG and Matroska encapsulate a very large variety of video+audio codecs, with Matroska being easier to expand bugt OGG being more robust.

I downloaded Linux x86 version of 3.7a4webm. Opted into html5 video, and used this search http://www.youtube.com/results?search_query=web&webm=1 and tried to play videos such as http://www.youtube.com/watch?v=BrXPcaRlBqo but it never loads. I just get
loading "throbber" and it says webm below the video.



It's definitely not attempting to load the x264 version as that yields



Opera beta's webm not working for me, not even their example. Also their implementation of webm is not compatible with Firefox's (youtube gives me the latter screenshot on a webm video and firefox can't view opera's example at all).

What doesn't make sense is container formats. "This device supports AVI", right which codecs? And you need matching codecs at that. It's all fine until you find yourself needing to convert videos. MKVs in my experience require more processing power than other containers even at comparable resolution. MKV includes subtitles within the container, otherwise I don't see the appeal other than in the bluray ripping community.

I already converted a lot of videos to OGG so maybe I'm just complaining that they want to switch formats - arbitrarily.

Of course it's not so much arbitrarily as $$$. Flash gets to live on by embedding vp8 within itself. Plugins now have plugins.

Maybe if they get this working and with less resources than Flash, then I'll change my mind.

Oh and as far as hardware support goes, they mean hardware decoding as opposed to software. There are chips that decode MPEG and chips that decode mobile JAVA for instance. It's an advantage on both embedded devices as well as high-end gaming platforms.

In Mozilla's response to H.264, they claimed they had done their homework pretty well & it lead them to not support H.264. So it seems pretty significant that they've joined in here. Maybe Google said it would protect them if they got sued b/c of VP8. All for one, one for all, like The Three Musketeers.

Some interesting updates from xiph.org regarding Theora, in particular ARM and OMAP DSP optimisations funded by Google and Mozilla respectively.

I wonder why Google would be funding a Theora build under BSD when they have just opened VP8 under BSD, unless it's not Theora but VP8? And Mozilla sponsoring OMAP DSP optimisations?

As for Opera with WebM support - worked fine for me on Win7-64.

What is that supposed to mean? Flash's support for VP8 is no more a plugin-in-a-plugin than its support for h.263/Sorenson Spark, VP6, h.264, etc.

But of course Flash wil adapt to keep alive -- if this hadn't happened, and Mozilla's delusions of shifting the entire web to Theora had actually worked, do you think they would not add Theora support?

This isn't about making Flash wither and die (at least not directly) -- it's about removing one of the biggest obstacles which was preventing HTML5 video from even competing -- lack of a decent codec with near-universal presence. VP8 has quality comparable to h.264 for reasonable bitrates, unlike Theora. Like Theora, it beats h.264 at low bitrates (such as for video chats), and it still seems to get beaten at high enough bitrates, but the crossover point is significantly higher now, making youtube etc. conversion feasible without adding bandwidth.

Very few embedded devices have dedicated MPEG chips, and I don't even know of any dedicated solutions for MPEG 4 codecs, although I suppose there may be some. These days it's very much DSPs everywhere; the only question is how many embedded devices lack an easy ability to upload new DSP code for additional codecs. But most such devices wouldn't be able to browse the web usefully anyway, so it doesn't seem like a big deal for a web video standard....

So what do we need? A vp8 codec that can use the OMAP3 DSP?

http://e2e.ti.com/blogs_/b/mobile_mo...open-arms.aspx

Yeah for sure - so OMAP4 will have direct support for VP8, but I would have thought a codec for vp8 could get some assistance from the DSP that is already in the n900?

What does the quote from TI mean? To me, it means nothing - not unless they commit their source code/support to upstream projects.

Having a library of suitable codecs (as binary blobs?) from TI is wonderful and all, but pretty useless unless they end up in an actual product. Remember, TI had 3D drivers for OMAP2 yet they benefited nobody other than the marketing team who could say, genuinely, that 3D is fully supported by TI.

I'd take the comments from TI with a large pinch of salt.

Given the way Theora can now run on the n900's DSP, couldn't a developer do the same for VP8? Granted this is a suboptimal solution, optimal being that it comes in an official Nokia update, but it could do the trick.