Steve Jobs says no to Google's VP8 WebM codec
In reply to a email asking his thoughts on Google's announcement of the royalty-free WebM video codec, Steve Jobs reportedly simply forwarded back the critical expose profiled yesterday by AppleInsider.
Jobs' terse reply to the question "What did you make of the recent VP8 announcement?" left little doubt about Apple's views on the matter, because the linked report Jobs returned, written by x264 developer Jason Garrett-Glaser, castigated the new specification for video compression as being weak, incomplete, and undoubtedly encumbered by patent issues.
Note that x264 is an open source project for encoding H.264 compliant video. It has no inherent bias in promoting the H.264 specification over VP8. Garrett-Glaser has been a vocal critic of elements of the MPEG standards process, the x264 project itself is an effort to get around commercial licensing of MPEG's technologies, and the similarities between VP8 and H.264 mean that x264 could likely be adapted to encode VP8 as well.
Jobs' problem with VP8 is not ideological
Critics of Apple have settled on condemning the company for not rushing to embrace the free new video codec, which Google acquired as part of the technology portfolio of On2 and has published as an open source, royalty free specification. However, the only real value of VP8 over H.264 is that it is royalty free; it's not as sophisticated as the latest MPEG-4 specification in terms of compression quality or efficiency, nor does it offer a comparable technical range of use.
VP8 can't currently support Apple's needs as a mobile-optimized codec for its iPods, iPads and iPhones, nor is it suitable for high definition video encoding. VP8 is targeted directly at the web, where Google, Mozilla and other partners hope to use it to deliver video without the royalties required by H.264. Apple has no issue paying royalties to license MPEG's H.264 technologies because the royalties are not very expensive. The H.264 licensing fees are really only a relevant cost issue for groups like Mozilla wanting to give their software away for free.
The problem is that video experts such as Garrett-Glaser are reporting that VP8 is not only unfinished and incomplete, but will also run afoul of the broad range of patents covering the latest video compression and decoding technologies. Those patents are held by a wide consortium of vendors who have pooled their technology together under the direction of the ISO's Motion Picture Experts Group.
Apple is not the patent holder to worry about
Most reports on the issue have focused on Apple's membership in the MPEG coalition in order to write up a fashionably dramatic tale of conflict between Apple and Google on the issue of the VP8 codec.
However, the reality is that the patents related to H.264 are held by a large number of companies scaling from embedded chip makers to video software developers to university research groups to Blu-Ray hardware vendors. Apple has a minority stake in the MPEG-4 patent pool related to video codec technology, and would actually benefit from a sophisticated yet free video codec, were one to exist.
Rather than Apple, it is the actual codec patent holders that will come after VP8 looking for royalties, and unless Google subsidizes the world's use of VP8 (and it will not), it appears certain that anyone who uses the "royalty-free" codec will end up having to pay royalties for its use anyway. Because it is not immune to patent royalties, all VP8 really offers in comparison to H.264 is less sophistication and a more incomplete and immature specification.
A strangely ideological puzzle
The apoplectic fit surrounding H.264 and its patented technology is quite bizarre in that the MPEG-4 working group that developed it didn't invent some new patent problem that hasn't existed before and desperately needs to be solved.
Those insisting on calling H.264 "proprietary" because it is not completely free have never referred to other open, but not free standards (including the MP3 audio format), as "proprietary." Additionally, it was Apple's iTunes that spearheaded the move from MP3 toward the more modern and efficient AAC audio format in order to shift the world's technology from the more expensive to license MP3 format to the cheaper and/or free to use AAC. If there were a cheaper, suitable option to paying for H.264, Apple would similarly jump on it.
Back in the early 2000s, both Microsoft and Sony were pushing to replace MP3 with completely proprietary formats: Microsoft's Windows Media Audio and Sony's ATRAC. None of the current critics of H.264 in the mainstream tech media had any issue with that potential shift toward completely proprietary and closed standards back then. It was only the success of the iPod that drove the world toward AAC as an open standard and preserved MP3 as a compatible fallback.
The web wants to be free
The web presents a slightly different set of circumstances because open source projects like Mozilla's Firefox can't legally bundle support for H.264 without paying royalties on the technology. It would be ideal if the world had video playback technology that was competitive, efficient for use on mobile devices, and completely free.
Unfortunately, this simply does not exist in the world of video where massive amounts of research and development go into developing new technology to encode, transmit and decode video efficiently. That results in researchers and scientists needing to get paid for their efforts.
The last time the web's media was significantly encumbered by a patent issue was with the GIF image file format, which was not itself patented, but which used a compression technology that Unisys claimed to own in the mid 90s, long after GIF had entered wide use on the Internet. The company vilified itself by demanding significant royalties from all companies that used or created GIF images commercially.
An effort to replace the patent-encumbered GIF with the new PNG format was initiated, but despite the creation of PNG as a technically superior, royalty free, and patent free new format, it was nearly impossible for the world to switch away from using GIF. While work on PNG started in 1995, it was not officially recognized as an ISO standard until 2003, when the patents on GIF finally began to expire (making the GIF problem a non-issue). Microsoft's Internet Explorer didn't effectively support PNG until 2005.
PNG is a relatively simple image format, dramatically less complex than the technology needed to encode, deliver and decode high quality video over an unreliable transport network like the Internet. That makes PNG's decade long march toward replacing GIF a cautionary tale for those expecting VP8 to serve as an immediate drop-in replacement for H.264, which has become the established video format on the web, for mobiles, and for a wide range of other uses ranging from video games to Blu-Ray.
Is VP8 patent encumbered?
Like GIF and its encumbrance with a compression-related patent, VP8 is said to be based closely enough upon H.264 so that the MPEG Licensing Authority will demand patent royalties for it. MPEG LA is already investigating and expects to set up a licensing program for collecting VP8 royalties, according to a report by All Things Digital's John Paczkowski.
Unlike PNG, VP8 is not technically superior to H.264, making it that much more difficult to propagate across the Internet as a new video standard, even if it were not patent encumbered at all. And unlike the GIF patents, the patents on H.264's brand new technologies won't expire for a long time.
Google maintains that it has reviewed the VP8 technology it acquired and is sure that nobody will have to pay patent royalties for the technologies it uses. However, Google's acquisition of On2 was rushed through in January, and barely six months later the firm's existing VP8 code implementation is being pushed out as a "specification," much the same way Microsoft rushed its Office file formats through a standards process to avoid having to support the already formally standardized OpenOffice formats.
Google was demonstrably in a hurry to quickly deliver a fix for the issue Mozilla raised about the use of H.264 on the web just last summer. In contrast, Apple, Netscape and Sun all spent several years working through the code for projects they planned to release as open source (including Darwin, Mozilla, and OpenSolaris) in an effort to avoid any patent-related surprises. No project has ever raced code from fully proprietary to open source in six months, casting serious questions on the quality of Google's supposed vetting process for VP8.
It's therefore no surprise that Microsoft is saying it won't bundle support for VP8 with Internet Explorer, even though the code is already available, and that Jobs is intimating Apple's thumbs down for the H.264 alternative Google is publishing, but not offering to indemnify from patent attacks.
Jobs' terse reply to the question "What did you make of the recent VP8 announcement?" left little doubt about Apple's views on the matter, because the linked report Jobs returned, written by x264 developer Jason Garrett-Glaser, castigated the new specification for video compression as being weak, incomplete, and undoubtedly encumbered by patent issues.
Note that x264 is an open source project for encoding H.264 compliant video. It has no inherent bias in promoting the H.264 specification over VP8. Garrett-Glaser has been a vocal critic of elements of the MPEG standards process, the x264 project itself is an effort to get around commercial licensing of MPEG's technologies, and the similarities between VP8 and H.264 mean that x264 could likely be adapted to encode VP8 as well.
Jobs' problem with VP8 is not ideological
Critics of Apple have settled on condemning the company for not rushing to embrace the free new video codec, which Google acquired as part of the technology portfolio of On2 and has published as an open source, royalty free specification. However, the only real value of VP8 over H.264 is that it is royalty free; it's not as sophisticated as the latest MPEG-4 specification in terms of compression quality or efficiency, nor does it offer a comparable technical range of use.
VP8 can't currently support Apple's needs as a mobile-optimized codec for its iPods, iPads and iPhones, nor is it suitable for high definition video encoding. VP8 is targeted directly at the web, where Google, Mozilla and other partners hope to use it to deliver video without the royalties required by H.264. Apple has no issue paying royalties to license MPEG's H.264 technologies because the royalties are not very expensive. The H.264 licensing fees are really only a relevant cost issue for groups like Mozilla wanting to give their software away for free.
The problem is that video experts such as Garrett-Glaser are reporting that VP8 is not only unfinished and incomplete, but will also run afoul of the broad range of patents covering the latest video compression and decoding technologies. Those patents are held by a wide consortium of vendors who have pooled their technology together under the direction of the ISO's Motion Picture Experts Group.
Apple is not the patent holder to worry about
Most reports on the issue have focused on Apple's membership in the MPEG coalition in order to write up a fashionably dramatic tale of conflict between Apple and Google on the issue of the VP8 codec.
However, the reality is that the patents related to H.264 are held by a large number of companies scaling from embedded chip makers to video software developers to university research groups to Blu-Ray hardware vendors. Apple has a minority stake in the MPEG-4 patent pool related to video codec technology, and would actually benefit from a sophisticated yet free video codec, were one to exist.
Rather than Apple, it is the actual codec patent holders that will come after VP8 looking for royalties, and unless Google subsidizes the world's use of VP8 (and it will not), it appears certain that anyone who uses the "royalty-free" codec will end up having to pay royalties for its use anyway. Because it is not immune to patent royalties, all VP8 really offers in comparison to H.264 is less sophistication and a more incomplete and immature specification.
A strangely ideological puzzle
The apoplectic fit surrounding H.264 and its patented technology is quite bizarre in that the MPEG-4 working group that developed it didn't invent some new patent problem that hasn't existed before and desperately needs to be solved.
Those insisting on calling H.264 "proprietary" because it is not completely free have never referred to other open, but not free standards (including the MP3 audio format), as "proprietary." Additionally, it was Apple's iTunes that spearheaded the move from MP3 toward the more modern and efficient AAC audio format in order to shift the world's technology from the more expensive to license MP3 format to the cheaper and/or free to use AAC. If there were a cheaper, suitable option to paying for H.264, Apple would similarly jump on it.
Back in the early 2000s, both Microsoft and Sony were pushing to replace MP3 with completely proprietary formats: Microsoft's Windows Media Audio and Sony's ATRAC. None of the current critics of H.264 in the mainstream tech media had any issue with that potential shift toward completely proprietary and closed standards back then. It was only the success of the iPod that drove the world toward AAC as an open standard and preserved MP3 as a compatible fallback.
The web wants to be free
The web presents a slightly different set of circumstances because open source projects like Mozilla's Firefox can't legally bundle support for H.264 without paying royalties on the technology. It would be ideal if the world had video playback technology that was competitive, efficient for use on mobile devices, and completely free.
Unfortunately, this simply does not exist in the world of video where massive amounts of research and development go into developing new technology to encode, transmit and decode video efficiently. That results in researchers and scientists needing to get paid for their efforts.
The last time the web's media was significantly encumbered by a patent issue was with the GIF image file format, which was not itself patented, but which used a compression technology that Unisys claimed to own in the mid 90s, long after GIF had entered wide use on the Internet. The company vilified itself by demanding significant royalties from all companies that used or created GIF images commercially.
An effort to replace the patent-encumbered GIF with the new PNG format was initiated, but despite the creation of PNG as a technically superior, royalty free, and patent free new format, it was nearly impossible for the world to switch away from using GIF. While work on PNG started in 1995, it was not officially recognized as an ISO standard until 2003, when the patents on GIF finally began to expire (making the GIF problem a non-issue). Microsoft's Internet Explorer didn't effectively support PNG until 2005.
PNG is a relatively simple image format, dramatically less complex than the technology needed to encode, deliver and decode high quality video over an unreliable transport network like the Internet. That makes PNG's decade long march toward replacing GIF a cautionary tale for those expecting VP8 to serve as an immediate drop-in replacement for H.264, which has become the established video format on the web, for mobiles, and for a wide range of other uses ranging from video games to Blu-Ray.
Is VP8 patent encumbered?
Like GIF and its encumbrance with a compression-related patent, VP8 is said to be based closely enough upon H.264 so that the MPEG Licensing Authority will demand patent royalties for it. MPEG LA is already investigating and expects to set up a licensing program for collecting VP8 royalties, according to a report by All Things Digital's John Paczkowski.
Unlike PNG, VP8 is not technically superior to H.264, making it that much more difficult to propagate across the Internet as a new video standard, even if it were not patent encumbered at all. And unlike the GIF patents, the patents on H.264's brand new technologies won't expire for a long time.
Google maintains that it has reviewed the VP8 technology it acquired and is sure that nobody will have to pay patent royalties for the technologies it uses. However, Google's acquisition of On2 was rushed through in January, and barely six months later the firm's existing VP8 code implementation is being pushed out as a "specification," much the same way Microsoft rushed its Office file formats through a standards process to avoid having to support the already formally standardized OpenOffice formats.
Google was demonstrably in a hurry to quickly deliver a fix for the issue Mozilla raised about the use of H.264 on the web just last summer. In contrast, Apple, Netscape and Sun all spent several years working through the code for projects they planned to release as open source (including Darwin, Mozilla, and OpenSolaris) in an effort to avoid any patent-related surprises. No project has ever raced code from fully proprietary to open source in six months, casting serious questions on the quality of Google's supposed vetting process for VP8.
It's therefore no surprise that Microsoft is saying it won't bundle support for VP8 with Internet Explorer, even though the code is already available, and that Jobs is intimating Apple's thumbs down for the H.264 alternative Google is publishing, but not offering to indemnify from patent attacks.
Comments
No HD capability google left that little detail out. I should say poor HD capability.
Why not just buy out MPEG LA, release H.264 for widespread use and be done with it?
Because it isn't possible to buy out MPEG LA without buying all the member companies.
I don't get the problem with Mozilla supporting HTML5 video though. They don't need to buy a license, they just need to piggyback on the operating systems license. The only free software licensing issue that exists is with free operating systems like Linux or FreeBSD. Those users would be stuck with Flash for video unless somebody (Adobe maybe?) releases an HTML5 video plugin for free under their name or Mozilla changes their stance and licenses h264.
Maybe I don't fully understand the licensing, but it seems like it is flawed. Wouldn't someone just have to release a h264 library that was multi-platform and they could get away with buying one license that is used by multiple companies?
Because it isn't possible to buy out MPEG LA without buying all the member companies.
Any idea of how much the royalty stream is per year?
Wow Ie won't support it? Then it's dead before it even starts.
No HD capability google left that little detail out.
I'm sorry, but what?
Microsoft said IE will support it in that it'll play videos in IE as long as the codec is installed on the system. They're not shipping IE with support but it'll be a one-time one-click setup thing to install a codec to play it on the system.
Of course the codec has HD capability. In fact, the very first thing Google did with WebM was convert all HD YouTube videos to HD WebM. If you download a WebM capable browser like the latest Firefox WebM builds and you're in the YouTube HTML5 beta, you are looking at HD WebM.
To the author of the article:
The slant in this news story, and many on this site, is astonishing. Why do people feel the need to defend Apple in everything? They're a business with their own vested interests, they're in this to make money and protect their interest...not to be a humanitarian company. Try being more being more objective. This site reads like a state newspaper in China or Pravda. Apple's side of the story is presented in detail, then a token reference to the other side is made followed by a slew of opinions stated as fact to discredit them. You guys can do better.
Because it isn't possible to buy out MPEG LA without buying all the member companies.
I don't get the problem with Mozilla supporting HTML5 video though. They don't need to buy a license, they just need to piggyback on the operating systems license. The only free software licensing issue that exists is with free operating systems like Linux or FreeBSD. Those users would be stuck with Flash for video unless somebody (Adobe maybe?) releases an HTML5 video plugin for free under their name or Mozilla changes their stance and licenses h264.
You seem to be confusing things.
1) HTML5 video support is not a video codec in and of itself.
2) A plugin is not required for HTML5 video playback to work.
Mozilla supports HTML5 video. They don't support the H.264 codec, but rather only Ogg Theora.
You seem to be confusing things.
1) HTML5 video support is not a video codec in and of itself.
2) A plugin is not required for HTML5 video playback to work.
Mozilla supports HTML5 video. They don't support the H.264 codec, but rather only Ogg Theora.
HTML5 video isn't a codec, but the only supported codec most people care about is h264. If Mozilla only supported Ogg Theora, almost all HTML5 video wouldn't play in Firefox.
The plugin or library would just to get around licensing issues on Linux. It would not be needed on OS X or Windows.
You seem to be confusing things.
1) HTML5 video support is not a video codec in and of itself.
2) A plugin is not required for HTML5 video playback to work.
Mozilla supports HTML5 video. They don't support the H.264 codec, but rather only Ogg Theora.
Mozilla supports Ogg Theora AND WebM.
I am sorry but I did not read the entire article because it it did not deliver on the title within the first several paragraphs. I am becoming very disappointed in the recent journalism on this site.
Steve Jobs says no to Google's VP8 WebM codec
in other shocking news, the sky is blue.
Of course the codec has HD capability. In fact, the very first thing Google did with WebM was convert all HD YouTube videos to HD WebM. If you download a WebM capable browser like the latest Firefox WebM builds and you're in the YouTube HTML5 beta, you are looking at HD WebM.
Please point me to the page detailing this "fact".
You said that they "converted" all HD YouTube videos to HD WebM. That would be a huge task with very little benefit other than to demonstrate the VP8 WebM codec. I don't see the reason for this.
Mozilla supports Ogg Theora AND WebM.
But is Mozilla about losing market share if h264 really takes off (like it appears it may). If they lose market share then they lose search bar revenue and they could go out of business. They can probably hold off for a while, but they may get to the point where they financially can't afford not to license h264 or at least make an exception to support it for OS X and windows.
Please point me to the page detailing this "fact".
You said that they "converted" all HD YouTube videos to HD WebM. That would be a huge task with very little benefit other than to demonstrate the VP8 WebM codec. I don't see the reason for this.
Are you serious? It was explicitly mentioned in the keynote and it's live now. You can try it yourself, I just told you how.
Of course they did it. It's a perfect test case for their product before they release it.
All YouTube HD videos are already converted to WebM and Google is already in the process of converting ALL YouTube videos in the back catalogue to WebM, and ALL NEW videos are also converted into WebM immediately.
"Little benefit"? Honestly? Promoting their own standard, testing their own code, "little benefit"?
The question seems to now be why did Google decide on the VP8 WebM codex?
Great article. My only though is ... YouTube and Chrome. YouTube is the 800 pound gorilla where video content is concerned and if Google switched to using only VP8 WebM on YouTube and if Chrome supports it and IE and Safari don't maybe this is a Google trojan horse for Chrome's adoption. That sentence was way too long ...
I hereby proclaim the fore mentioned products should be know as the I P 3 group ( or the IP Cubed group) or product group or product line.
This should allow us to speed up all conversation and reference to this operating system / product family, while allow all of us to have some jargon that satisfies our need for an "insider catch phrase" that only we understand. Sort of like a gang sign for the rest of us, without having to be beat in or beat out.
Whud da ya say?
in other shocking news, the sky is blue.
How deep
I'm sorry, but what?
Microsoft said IE will support it in that it'll play videos in IE as long as the codec is installed on the system. They're not shipping IE with support but it'll be a one-time one-click setup thing to install a codec to play it on the system.
Of course the codec has HD capability. In fact, the very first thing Google did with WebM was convert all HD YouTube videos to HD WebM. If you download a WebM capable browser like the latest Firefox WebM builds and you're in the YouTube HTML5 beta, you are looking at HD WebM.
To the author of the article:
The slant in this news story, and many on this site, is astonishing. Why do people feel the need to defend Apple in everything? They're a business with their own vested interests, they're in this to make money and protect their interest...not to be a humanitarian company. Try being more being more objective. This site reads like a state newspaper in China or Pravda. Apple's side of the story is presented in detail, then a token reference to the other side is made followed by a slew of opinions stated as fact to discredit them. You guys can do better.
Explain to me why developers would be interested in this if apple or Microsoft won't bundle. Qiuicktime and media player and flash is already solving the problem. As to HD issue I assume writer knows what he is talking about. Sounds like the technology is not suited to large formats.
This common reference to Ipod -I phone - I pad is getting grammatically boring:
I hereby proclaim the fore mentioned products should be know as the I P 3 group ( or the IP Cubed group) or product group or product line.
This should allow us to speed up all conversation and reference to this operating system / product family, while allow all of us to have some jargon that satisfies our need for an "insider catch phrase" that only we understand. Sort of like a gang sign for the rest of us, without having to be beat in or beat out.
Whud da ya say?
Or just write iPhone OS.
Explain to me why developers would be interested in this if apple or Microsoft won't bundle. Qiuicktime and media player and flash is already solving the problem. As to HD issue I assume writer knows what he is talking about. Sounds like the technology is not suited to large formats.
The author demonstrably does not know what he's talking about, as it's live now and it looks great. The builds are available, download them and try it. Stop taking this propaganda for truth. Try it out.
And no, QT, MP, and Flash solve the problem of putting video online. But the video exists inside a container and can't be properly interacting with the web document. Plugins are a workaround, HTML5 video is a true solution.
Developers are interested in VP8/WebM it because the encoders and decoders are and will continue to be free and the quality is very good. Developers are interested in HTML5 Video because it lets you do a ton of cool things (interact with the page's DOM, use Javascript, CSS, etc).
I'm building an HTML5 video portal right now at work.
No, WebM is not AS good as h264 in terms of quality but it's about 95% of the way there. And yes, the royalty uncertainty with h264 as of Jan 1st, 2016 makes a lot of people nervous. My employer is using h264 right now but out of necessity -- they are nervous about it as we're becoming a vendor for video which can get very expensive very quickly. If the royalties arrive in 2016, we'll have a massive back catalog of h264 that'll cost us a fortune in annual fees. WebM is looking very, very good to them right now. There's a reason why everyone's jumping all over it unless they're already H264 licensees (Apple, MS).