X264 developer says Google's new VP8 WebM codec is a mess - Page 2

I fully understand the significance for web video and the future of the internet.

The problem is this reads like an attack article: Google dared come out with a rival to the iPhone so lets start posting negative articles about them.

Hm, maybe because you want to read it like that? Both formats are open. H.264 is superior and widely supported by hardware decoders (less CPU load, less energy consumption) and already the recording format on a lot of devices (camcorders, cameras, mobile phones, HDD video recorders, PVRs). Why add a lesser format that will only add complexity (record H.264 on my point and shoot, convert it to VP8/webm, then play it back in a browser without direct hardware support, so let the CPU do all the work a second time)... it makes little, if any, sense. It is a solution for which there is no need. Like a kid riding the bike without holding the handle bar: look Ma, no teeth... It is not better, only different.

I agree. This is meaningless to Apple. Apple can support VP8 just as easily as h264.

On a desktop/laptop maybe. On an iPhone/iPod/iPad: no. H.264 is supported by hardware decoders, Motion-JPEG (the other format supported on the iPad, and maybe the next iPhone) uses no complex compression (files are rather huge, but there is not a lot of work for the CPU). Adding VP8 support would decrease battery life. Do you think Apple will introduce the next iPad and say: now 5 hours of video playback (instead of ten)? I do not see this happening at all.

very true.

add to this that MPEG LA has waived the licensing fees for pretty much everyone until 2016 (and could waive them again or even drop them) and it is no shock that H.264 is so hot on the web etc

"X264 developer says Google's new VP8 WebM codec is a mess"..... And in other news, Flash developer says HTML5 is a mess.

Within hours of the announcement. everyone is rushing to spread FUD about VP8 and come to H264's defense. A surefire way to keep Flash relevant for years to come. Adobe must be delighted. VP8 is the best way to kill Flash by avoiding a stalemate in the video codec wars.It's a little premature to declare it unfit to replace H264 as the standard for video encoding.

If all people did was watch video that might be an issue. I think video is a small fraction of the time people spend on the iPhone, so it's not a big issue there. I don't think the iPad is particularly good for video either - it's 9.7" SD screen becomes a lot smaller when watching widescreen video.

Well, I was flying back from Washington DC to Frankfurt yesterday and there were 8 iPads in business class (plus mine). Absolutely everybody was watching video whenever I passed them. I think 720p video on the iPad looks absolutely amazing and it is definitely as big as watching my 52" LCD from a normal distance (12" or so) at home (maybe even bigger).

Anyhow, it is not a question of being an issue or not. Apple advertises video play back times - and 10 hours of video IS a selling point compared to almost all laptops and netbooks. They won't allow that figure to go down and they will not support a "foreign" format, if it messes up their specs.

Well, he may have a bias but he is an expert in a technology that is pretty opaque to most people. (Me for certain) As such his observations and detailed comments can be looked at and checked for verification. A Windows programmer would be able to make a lot of interesting comments about the differences that OS X has and his opinions about them. And scientific results should be verifiable regardless of the bias of the checker.

We're not just getting a conclusion, we're getting a dissection with notes.

So why couldn't they simply point people to a H.264 plugin. I mean that's how Microsoft is going to treat WebM in IE9, so they can distance themselves from any potential patent problems. And that's how I fully expect Apple to treat WebM on the Mac. With all of the extensions in Firefox, you're telling me that they couldn't add plugins for codecs?

As for their ideology, what about other patent encumbered codecs like GIF use to be? Mozilla has always supported them without trouble or ideological trauma. Is this a case of "getting some religion" because of those experiences? I fully appreciate their stance, but I think they come off as not open to a compromise that still respect their ideals.

Exactly. So far, to counter they call him spreading FUD, instead of dissecting the Assembly code of H.264 and it's many algorithms and comparing them to V8.

Let's see them.

As a mechanical engineer and computer scientist I don't trust gossip from web sites--I want to see the research.

Okay firstly, do you have any specific refutations of Jason Garrett-Glaser's very technical review? Or are you just labeling it as FUD as a knee-jerk reaction? If you have any real expertise and can refute any of Jason's claims then I'm interested. Otherwise...

Secondly, how does this have anything to do with Flash? Flash can play both H.264 and WebM within its container. They just announced their support. Adobe is FINE with it.

Thirdly, this is the first time most people have been able to do more than speculate on VP8 and its suitability as a replacement. I don't think Jason was saying that it's totally unfit. He thinks it needs a lot of work, and fears that it will be targeted for patent violation. Google has declared that it's ready for use. So there will be no shakedown period. The spec is final. If Jason is in any way right then that's a problem.

Yeah, the H.264 codec is really only a problem for big companies who have to pay millions to integrate it and don't want to pay up. Like Mozilla.

Mozilla dev can simply forward the decoding on to a system decoder, like MPlayer or VLC.

I don't really see a switch from H.264 as there are way too many devices out there that hardware accelerate it. What Google could do is offer VP8 instead of Flash but alongside H.264. This will require people to upgrade to a non-IE browser if it's HTML 5 video.

I could "dissect" x264 and make it sound like crap too. If you do not have a detailed understanding of it, you would never know the difference. If a "neutral" party does the dissection it is worth a look, but this guy most certainly has a strong bias, which makes his points irrelevant unless verified or backed up with independent investigation.

I'm not arguing that - and have no way of knowing if VP8 is any good or not.

My point is that if VP8 is any good, there's nothing to stop Apple from supporting it. If h264 is better, Apple can support that. Apple can support both, if they wish. If Billy Bob's Magic Software Coding comes out with another codec that's better than either, Apple can support that. Or all of them.

It just doesn't matter to Apple one way or another.

http://arstechnica.com/web/news/2010...m_campaign=rss

There is literally no excuse for Apple not to embrace WebM. Then again, there's no excuse for Apple not fixing their 5 year old H.264 video export color conversion bug, either, but that simple fix has yet to come.

The H.264 problem is well documented. AI readers: please read up on it.

I read the article and it was showing the perspective from a person who had been critical of Apple's H.264 and was stating his opinion on Google's VP8 WebM and its user-ability at the moment compared to H.264, which this person was not overly praising ( I mean the H.264)

Personally as long as VP8 WebM can be supported and does not interfere with Apple's development plans for future iPhone/OS improvements, then it should be no skin off Apple's back. If on the other hand it is a Adobe approach and half hearted attempt to support iPhone, then NO please.

Hence why I said, "And scientific results should be verifiable regardless of the bias of the checker."

However I wouldn't place his bias as strong enough to poison everything he says. He is a developer in an open source project and a contributor to ffmpeg. And if you read his previous post on the subject when all he could do is speculate, he gives a pretty even-handed overview of the situation and points out all of the pragmatic points.

Flash, Google, VP8, and the future of internet video
http://x264dev.multimedia.cx/?p=292

There is a comment made by a member of Xiph (which have their opposite bias) that helps give some perspective on it. He basically says that "the devil is in the details" about the patent infringement, meaning that they are similar but subtly different. On the code and spec, he seems to agree so there you go.

http://xiphmont.livejournal.com/5023...135231#t135231

It would have been nice if Google had opened it up for the video developers to help clean things up, but that's not how they decided to roll. The spec is frozen. Hey Google, what happened to being truly "open"?

Well, Adobe has stated that Flash will support WebM so it could just be the backend video that gets changed and Flash would take care of older browsers as well. And Microsoft has stated that IE9 will support it with a plugin. Of course that still means that all IE users will need to upgrade to the latest version or use Flash. Personally, I don't feel that Microsoft is pushing them hard enough and the web is hurting because of legacy support.

Google better keep supporting H.264. Even if WebM replaces H.264 in many use cases like YouTube, we're not there yet and they shouldn't force everyone to upgrade their devices right now to get there.

I agree sir. And in the interest of that debate, there is a comment from the "opposite" side at Xiph.

http://xiphmont.livejournal.com/5023...135231#t135231

He seems to say that the specs are similar but subtly different in the details, so wouldn't be infringing. On the spec being "imprecise, unclear, and overly short, leaving many portions of the format very vaguely explained" he replies:


And to give some perspective to this, Jason had almost two days to review VP8 and the people at Xiph had "weeks" and qualify that "Some design chunks look very similar even to the trained eye..." So I believe that Jason isn't practicing any intentional bias in this debate.

Seemingly, you know nothing about the products produced and sold by these companies.

You may think that is what you are doing, but from this site's perspective, you are coming here to become a product that they sell to their advertisers.

And stories about Google and Adobe move a lot of product.

I read the article to which AI is referring and that was not the interpretation I got. My read was that it is inferior to H.264, on par with several other specs, and far better than Theora (aka VP3). He also stated that there is plenty of room for improving the encoder/decoder, but that it would never be better than X264's implementation of H.264 for a couple of fundamental reasons. He also pointed out that the Open Licensing scheme is a strong positive in its favor, and that combined with the quality improvements over Theora, could make it a contender.

I think that it will do fairly well because of the objections Mozilla have with H.264. I don't think it will dominate though because of the potential for H.264 patent owners to sue over the similarities in the spec, and because MS has thrown its weight behind H.264 along with Apple.

Don't go to Google's websites. Problem solved. No more havoc.

That is precisely the model used by this website, AppleInsider.

They give away content to harvest info via cookies that they sell to advertisers. Essentially the Trojan Horse model.

But given that Google and AI are using the same "model", does that change your attitude towards the model?

You've gotta choose your side.

Steve has declared that Google is evil. Therefore, you cannot like anything about Google without automatically being disloyal to Apple.

Those are your choices. Deal with them.

Now you are just grasping at straws and bringing in different sources on different topics. The fact remains, h264 developer says competing VP8 is a "mess". Just like when Microsoft says Windows is better than OSX, or Faux News says they are Fair and Balanced.

You don't have to lie about Android. http://source.android.com/
It's free and open and public domain. Why do you think so many phones have it? Why do thing there can be so many alternate UI designs with the same OS? Open in this case, means you can change it and use it freely to suit your needs. Google will never sue you, even if you create an OS of your own and distribute as an alternate Android OS. You could call it... Robot OS.

AppleInsider needs to really clarify their definition of Open. Anything that is riddled with patents that you run the risk of being sued over just for using is not Open. It's just as bad as Apple. Honestly, I would challenge MPEG-LA to sue over their codec. Just like Henry Ford tore down the illegally monopolizing patent for Automobile Engines, Google would tear down the illegal monopolizing patents of MPEG-LA. How can you monopolize a mathematical algorithm anyway? Really our patent system is broken. Software should not be patentable. Copyrightable... Yes. Patentable. No. After call, software is authored just like a book. Else maybe we should start allowing books to be patentable. I guess that would mean those that write books about a wizard can get sued.

Really you need to understand the problem with using H.264. It may be cool and all, but it will not work unless it truly becomes Open and Free to use.

And really, the author doesn't know a thing about Mozilla. Firefox works just fine. It's even faster than Safari. Maybe not on Apple machines, but neither does Flash. Does that tell you something? Apple's OS truly sucks at handling code other than their own. No wonder developers hate programming for that OS. Apple never supported developers, and until it learns how, it will never get love from them. Microsoft learned that lesson long ago. Now they have the best developer support sites, second to none.

Mozilla also added tabbed browsing long before Safari and other browser did. Here is the wiki on Firefox: http://en.wikipedia.org/wiki/Mozilla_Firefox

Honestly, any reporter that reports false and opinionated tabloid information, such as this, should not be a reporter. Get your facts straight and subdue your opinions before writing such a stupid article. And don't rely on one idiots opinion. And stop twisting the truth to support your arguments. You are not doing the world a favor by spreading lies and partial truths.

Except that according to the article, VP8 is likely infringing on H.264 patents and that Google won't help anyone out who gets sued.

This truly open enough for you???

To submit your code contributions visit this page: http://www.webmproject.org/code/cont...tting-patches/

How much more open can you get?

Where was the rule written that open software could not be patented?

You have no point and is the reason you are all over the place with this. The mathematical algorithm itself is not patented. Its a combination of mathematical algorithms in sequence that perform a specific function. Why would a company invest millions of dollars in resources to create software if anyone is free to use their effort without having sacrificed anything. These companies are in business is to make money not give things away for free.

Software and media are completely different. Software actually performs an active function. A book or movie are passively read and watched, they perform no active function in of themselves.

Before rambling a copy/paste reply, read what was actually stated.

Android is open source, but the valuable part of android is the "with Google" apps, which are free (if you abide by Google's restrictions so that Google is making ad revenue from your use) but not open or free to use however the end user wants to adapt them.

Patches to a closed spec is equivalent to asking for free labor to repaint your house, landscape your existing yard, give your roof new shingles, etc. The specification was done before it was discussed.

Open is WebKit.

Software doesn't do anything unless a CPU interprets it. It's still just written instructions that has no function until it is executed by the little electrical pieces inside of your computer.

It's just like having a cookbook. It's just a set of instructions on a page until you actually follow those instructions.

Stop trying to make software into something that it is not.

How is a combination of algorithms more patentable versus one? Honestly. That's a dumb argument.

H.264 is about as open as Flash. Use it at your own risk.

Closed spec??? Uh.. Hullo? The code is right there. You have the specs. You can take the code, edit it, make changes and make your own codec and call it MyVideoCodec without the risk of being sued by google. Let's see you do that with H.264.

Open spec for H.264? Open as much as Flash is.

True. Google Apps are closed source.

But, as you confirmed, Android is open source and you can aways create your own apps. Your not forced to use Google Apps, nor are you tied down to Google. Yahoo could come out with their own Android platform if the wanted to. And so could you.

How many free cookbooks did Julia Child write?

Since when does open mean license free?

Boy did you entirely miss the analogy? How may patents are on Julia Child's cookbooks?

My point is software should not be patentable. Copyrightable, yes, but patentable, no. Consumers are losing because this is going on.

Free is not the issue. The patents that riddle the purported "Open" H.264 video format are the big elephant in the room. My point is, that because software is written much like a cookbook, instruction manual and just about any novel out there, it should not be allowed to have patents on it. This is what allows businesses to have illegal monopolizing patents, such as what surrounds the H.264 codec.

Did you know that Thumbnails are patented. It's patent infringement for you to use thumbnails on a web page. Apple, Google and Microsoft are being sued over it. How many more ridiculous patents should we as tax payers and consumers have to pay for.

Am I the only one here that thinks this is messed up?

Who said anything about a license? All open formats have licenses. Many very unrestrictive licenses.

So what is "Open" in the sense used in the article?

This?
http://en.wikipedia.org/wiki/Open_standard

This?
http://en.wikipedia.org/wiki/Open_specifications

Or this?
http://en.wikipedia.org/wiki/Open_format

Or this?
http://en.wikipedia.org/wiki/Open_content

Or this?
http://en.wikipedia.org/wiki/Open_source

Obviously there are degrees of Open, and I'm saying it's not truly open unless it's fully open.

H.264 is open just a crack. We have the specifications, but if someone were to use those specifications to create their own codec, they could get sued due to patents. Which, then tightly closes the door on the format.

Adobe has recently provided the Specifications for Flash, thus, it is open in the same sense. If I decided to take up the challenge, I could develop a Flash player using those specifications without using an ounce of code from Adobe.

Does that make sense? Consider what Apple is selling you out for. I personally want nothing to do with it.