Linked by Thom Holwerda on Wed 12th Jan 2011 17:44 UTC
Internet & Networking With yesterday's news that Google will be dropping H.264 support from the Chrome web browser, the internet was split in half. One one side, there's people who applaud the move, who are happy that Google is pushing an open, royalty-free and unencumbered video codec (irrespective of Google's motivation). On the other side, there are the H.264 supporters, who believe that H.264 is the one and only choice for HTML5 video. One of the most vocal and public figures in the latter group is John Gruber. Following his five questions for Google, here are ten questions for Gruber about WebM, H.264, and standards on the web.
Thread beginning with comment 457337
To read all comments associated with this story, please click here.
Hyperbole.
by FellowConspirator on Wed 12th Jan 2011 20:22 UTC
FellowConspirator
Member since:
2007-12-13

The article's a little hyperbolic and has some inaccuracies.

1.) strictly speaking, h.264 and WebM are equivalent with respect to royalties and patents with the exception one of the holders of VP8-related patents (Google) has given public license to them; WebM will be cheaper, but not free. The royalty-free nature of h.264 is now set to expire in 2014, but so will many of the patents covering it.

2.) MPEG-LA wouldn't be a patent-troll per se. They were formed by the consortium of the technology developers and original patent holders and they've always been up front about their claims.

3.) Is correct, but it's not clear why it's important. It will take a few years before the technology is pervasive, at which time it probably won't be relevant.

4.) Some browser makers. However, web browsers make a minority of the video distribution and production space. Almost all modern video tech outside of the web is based off h.264.

6.) VP3 is improperly compared to h.264. It's much closer to ISO/IEC 14496-2 which came out a year prior to VP3. ISO/IEC 14496-2 (h.264) are most closely comparable; they are contemporaneous and use many of the same technical approaches.

7.) This overlooks the fact that perceived patent issues with VP8 and it's predecessors contributed to the sale of VP8 to Google.

9.) It's a loaded question. The browser should rely on the platform it runs to provide support for codecs. The only reason this conversation exists at all is because some want to move responsibility for implementing video codecs directly into the browser itself instead of the operating environment (the idea being that it might simplify implementation and consistency).

10.) I don't think Apple cares, honestly. They use h.264 because everyone else does. Their platforms offer video playback and transcoding through a common API and a codec plugin architecture (as does Microsoft). Adding support for WebM really only requires pushing out the codec in an OS update (or an OS X download). You don't find WebM anywhere yet, thought, so why add bloat?

The arguments are kind of silly. Google has good reasons to push for adopting a codec wholly controlled by them as the de facto standard. It gives them the ability to be the sole arbiter on matters of DRM, etc. For Apple and Microsoft, it's pretty much an annoyance. For the video production and distribution industry, it'll be an enormous pain in the butt.

Reply Score: 4

RE: Hyperbole.
by WereCatf on Wed 12th Jan 2011 20:36 in reply to "Hyperbole."
WereCatf Member since:
2006-02-15

9.) It's a loaded question. The browser should rely on the platform it runs to provide support for codecs. The only reason this conversation exists at all is because some want to move responsibility for implementing video codecs directly into the browser itself instead of the operating environment (the idea being that it might simplify implementation and consistency).

Hooking to system APIs is fine for a browser aimed for that platform but it just simply doesn't suit a browser that is aimed at being consistent on multiple platforms. Like for example if Firefox hooked to system APIs one couldn't anymore say "works on Firefox", they'd have to say "works on Firefox if [condition]"

Secondly, it would indeed create lots of extra work. At the moment if there is a problem with one or another feature and they get a bug report then it's quite easy to find, but if they hooked to native APIs and there was a bug somewhere there Mozilla would get lots of bug reports that actually have nothing to do with Firefox. Ie. it would mean lots of bogus bug messages and more work to sift through all of them. Not to mention the work needed to support all the different APIs. It does all add up to quite a lot.

Reply Parent Score: 4

RE[2]: Hyperbole.
by steve_s on Thu 13th Jan 2011 09:52 in reply to "RE: Hyperbole."
steve_s Member since:
2006-01-16

Hooking to system APIs is fine for a browser aimed for that platform but it just simply doesn't suit a browser that is aimed at being consistent on multiple platforms. Like for example if Firefox hooked to system APIs one couldn't anymore say "works on Firefox", they'd have to say "works on Firefox if [condition]"


And yet every single cross-platform browser hooks into system APIs for many other aspects of their functionality, such as accessing files, drawing windows, etc. This is done through shim libraries.

Secondly, it would indeed create lots of extra work. At the moment if there is a problem with one or another feature and they get a bug report then it's quite easy to find, but if they hooked to native APIs and there was a bug somewhere there Mozilla would get lots of bug reports that actually have nothing to do with Firefox. Ie. it would mean lots of bogus bug messages and more work to sift through all of them. Not to mention the work needed to support all the different APIs. It does all add up to quite a lot.


Yes, it would create some extra work, but I'm not convinced it would be all that much. All that's required for each platform is a fairly thin shim library that converts to the OS provided media framework. The functionality of all media frameworks is largely equivalent and generally very similar in functionality.

In contrast, cross platform game developers write engines that use either DirectX or OpenGL for rendering - the differences between how those two libraries work, especially in their latest incarnations, is an order of magnitude more complex than media playback, yet they still manage just fine.

As for bug reports, the crash logs would clearly indicate where the error occurred.

Reply Parent Score: 2

RE[2]: Hyperbole.
by segedunum on Thu 13th Jan 2011 17:16 in reply to "RE: Hyperbole."
segedunum Member since:
2005-07-06

Hooking to system APIs is fine for a browser aimed for that platform but it just simply doesn't suit a browser that is aimed at being consistent on multiple platforms.

It also restricts new competitors coming in with new devices and new software systems, since they're obviously not going to have h.264 and will need to get it - if the oh, so open consortium of competitors decides to be nice.

Reply Parent Score: 2

RE: Hyperbole.
by galvanash on Wed 12th Jan 2011 21:32 in reply to "Hyperbole."
galvanash Member since:
2006-01-25

The article's a little hyperbolic and has some inaccuracies.


Not as many as your post...

1.) strictly speaking, h.264 and WebM are equivalent with respect to royalties and patents with the exception one of the holders of VP8-related patents (Google) has given public license to them; WebM will be cheaper, but not free. The royalty-free nature of h.264 is now set to expire in 2014, but so will many of the patents covering it.


That is based on your assumption that WebM is encumbered by patents that are not owned by Google. Google doesn't believe there are any. Also, that last sentence should read "The royalty-free nature of non-remunerated (i.e. free) content over the internet is now set to expire...", but nice attempt at trying to cast that net over the whole thing.

2.) MPEG-LA wouldn't be a patent-troll per se. They were formed by the consortium of the technology developers and original patent holders and they've always been up front about their claims.


They are run by one (Larry Horn) according to the linked article, which was the point being addressed I believe. I'll leave this one alone though as it doesn't make much difference to me.

3.) Is correct, but it's not clear why it's important. It will take a few years before the technology is pervasive, at which time it probably won't be relevant.


You will need to explain that one further if you want to convince me. MPEG-2 is still widely used after over 15 years, and while its use is waining there is no reason to believe webm won't be relevent for just as long.

4.) Some browser makers. However, web browsers make a minority of the video distribution and production space. Almost all modern video tech outside of the web is based off h.264.


So what? What exactly does that have to do with HTML5? I don't see how that is relevant at all.

6.) VP3 is improperly compared to h.264. It's much closer to ISO/IEC 14496-2 which came out a year prior to VP3. ISO/IEC 14496-2 (h.264) are most closely comparable; they are contemporaneous and use many of the same technical approaches.


Im not even going there. This whole "who had what first" line of reasoning is pointless - none of it means squat unless there is a lawsuit at some point. We'll see.

7.) This overlooks the fact that perceived patent issues with VP8 and it's predecessors contributed to the sale of VP8 to Google.


Show proof of that. I'm not saying your wrong, just that I have never seen evidence that On2 believed they were infringing.

9.) It's a loaded question. The browser should rely on the platform it runs to provide support for codecs. The only reason this conversation exists at all is because some want to move responsibility for implementing video codecs directly into the browser itself instead of the operating environment (the idea being that it might simplify implementation and consistency).


Total bull. Ive seen this argument so many times it makes my hair stand up. That whole line of reasoning is nothing more than a clever ruse to hide the fact that it isn't possible for an open source or free to the user product to implement h.264 legally - the license as it is currently defined simply doesn't have a provision in it that would even allow for such a product as it is explicitly defined by sales... So you say "Let the OS do it", which in reality means "you don't make money so you don't even exist to us - you don't matter".

The best place to implement the video stack for HTML5 is in the browser itself. Google does it this way, Opera does it this way, and Mozilla does it this way. IE and Safari have their reasons for choosing to use the system codecs, which are plainly obvious to anyone with a brain.

10.) I don't think Apple cares, honestly. They use h.264 because everyone else does. Their platforms offer video playback and transcoding through a common API and a codec plugin architecture (as does Microsoft). Adding support for WebM really only requires pushing out the codec in an OS update (or an OS X download). You don't find WebM anywhere yet, thought, so why add bloat?


Nice dodge. They care because:

1. They are part of the h.264 patent pool.
2. They use h.264 extensively and almost exclusively.
3. They are backers of a streaming standard that uses h.264 exclusively, and it is the only streaming method they officially endorse for use over 3G (HTTP Live Streaming)

In fact I would go as far as saying that "they don't care" is about the most dishonest thing you posted here - they absolutely care.


The arguments are kind of silly. Google has good reasons to push for adopting a codec wholly controlled by them as the de facto standard. It gives them the ability to be the sole arbiter on matters of DRM, etc. For Apple and Microsoft, it's pretty much an annoyance. For the video production and distribution industry, it'll be an enormous pain in the butt.


How is it "controlled" by them? I thought other people had patents on it ;) Regardless, it is BSD licensed, the minute the published that license they lost all control over it. For anyone wanting to write free software that encodes video for web distribution, it is the only game in town.

I'm not going to get into a pissing contest about one or the other of these formats winning or whatnot, its irrelevant. At worse WebM causes a minor dent in MPEG-LA's protection racket, it isn't going to have any affect on the broadcast world or most hardware devices, Apple will likely never adopt it so it will be relegated to Android for mobile, it will always be a red headed stepchild when it comes to Windows, etc. etc. I'm not naive enough to think that it matters at all in the grand scheme of things. But it does matter to people who do not want to get dirty hands from dealing with the likes of MPEG-LA.

Reply Parent Score: 8

RE[2]: Hyperbole.
by chrisr on Thu 13th Jan 2011 12:48 in reply to "RE: Hyperbole."
chrisr Member since:
2005-08-26

Google claim to believe that VP8 does not infringe on any third-party patents, but they don't offer indemnity to companies planning to use VP8, so apparently the belief is not held with much conviction.

If they released some analysis of their patent clearance check I might place more weight on their claim. There are a huge number of patents covering all conceivable aspects of a video codec, so I am extremely skeptical that VP8 does not infringe any third party patents.

Reply Parent Score: 1

RE[2]: Hyperbole.
by segedunum on Thu 13th Jan 2011 17:10 in reply to "RE: Hyperbole."
segedunum Member since:
2005-07-06

At worse WebM causes a minor dent in MPEG-LA's protection racket, it isn't going to have any affect on the broadcast world or most hardware devices, Apple will likely never adopt it so it will be relegated to Android for mobile, it will always be a red headed stepchild when it comes to Windows, etc. etc.

I'm not so sure. When you have a format like this that a major video site uses, a format that everyone uploading to there will use and a format that anyone has the freedom to implement these things have a habbit of encroaching like sand filling in a lot of crevices.

That's what the MPEG LA is worried about and why Gruber, Apple and Microsoft have ratcheted up the rhetoric. Once the pillars fall down that keeps the exclusive control of a club intact it doesn't have much of a future.

Reply Parent Score: 3

RE: Hyperbole.
by cheemosabe on Wed 12th Jan 2011 22:24 in reply to "Hyperbole."
cheemosabe Member since:
2009-11-29

1.) I think you are hyperbolic here. Won't even bother.

4.) You know why WebM is called as such? The discussion is about the future of the web.

9.) If you want your browser to use OS codecs use one that does. This whole discussion doesn't apply to you anymore.

10.) You're kidding, right? Nobody should even try to innovate then.

Reply Parent Score: 1

RE: Hyperbole. - #9.. it's for security
by jabbotts on Thu 13th Jan 2011 15:03 in reply to "Hyperbole."
jabbotts Member since:
2007-09-06

One reason for implementing support within the browser is security; the browser is a huge target and easy access point into the OS.

- if browsers feed content to the OS media framework a successful exploit (malformed media content) gains access to the entire OS. Play exploit against vuln in the OS framework and you pwn the OS.

- if browsers play content internally a successful exploit may be contained within the browser's own sandbox; it does not rely on things outside it's own security sandbox. Play exploit against vuln in the browser and you pwn the browser, not the OS.

IE had a vuln in the system wide ANI related library; any program that displayed an animated mouse pointer was exploitable including IE. Use a malicious mouse pointer within your website and you would be able to feed stuff directly through IE into Windows.

Exploits against Flashplayer have also allowed people to break out of the browser into the greater OS. Flashplayer runs externally from the browser so bugs in it affect more than just the browser.

Reply Parent Score: 2

RE: Hyperbole.
by PresentIt on Fri 14th Jan 2011 22:19 in reply to "Hyperbole."
PresentIt Member since:
2010-02-10

1) H264 is closed. WebM is open.
2) They are patent trolls.
4) Web browsers are needed to view video on the web, which is the context here. Pay attention.
7) Quit it with the FUD. No actual patent claims have been filed for VP8.

WebM is not controlled by Google. You Apple fanboys need to stop lying.

Reply Parent Score: 2