Sorry, I read that too quickly. You were talking about a browser implementation and not a content implementation, I think.

A bit of both.

Asa Dotzler, on July 15, 2009 at 5:39 pm, said:

Google’s been pushing Chrome not just on YouTube, but with a giant and first of a kind banner ad on the front page of Google.com for months now. The banner ad has a big Chrome log, a big “Install Google Chrome” button and the message “A faster way to browse the web”. It’s had only a marginal impact on Chrome’s growth and couldn’t even lift Chrome up to Firefox’s early growth rate.

I’ve seen it when testing in IE6. Very, uh, subtle. But given the difference even one word can make in a call to action, I’m optimistic about a “make your video suck less” banner’s appeal – because it can promise a tangible difference to something the user cares about, right now. Still, selling a product to someone who has no idea they need it isn’t easy, and it’s a situation you’ve obviously had more experience of than me; so I’ll concede that such optimism is probably misplaced.

Sorry, I read that too quickly. You were talking about a browser implementation and not a content implementation, I think.

Sure. We probably don’t know which is true yet. I’m leaning pretty far to one side in my predictions, though.

I don’t think anything gets lots of people switching to HTML5-capable browsers. There are no quick wins when it comes to moving the 1.whatever billion web users in a new direction unless you own the Operating system that most of them use. Even then it’s years before you get them moved to your latest version.

It’s a long slog and we’ve got some more years before the modern browsers, in total, pass IE’s best offering globally. (note, in some places in the world, that’s already happened. but it’s got a ways to go before it’s solidified across a range of geographies.)

I’m of the opinion that a kick-ass <video> implementation isn’t going to have _any_ measurable impact on migration from other browsers. I’ll re-visit this post in a couple of years and admit if I’m wrong. But I don’t think I will be.

So there are codecs that give 10x the compression for the same quality as today’s JPEGs? If it’s 10 times better, surely Nikon and Canon are shipping it in all their most popular cameras. Right?

Why would they when anything but JPEG would likely frustrate the average consumer? It’s the exact same situation with the img tag – in order for it to be widely implemented there has to be demand from people using it, but people won’t use it until there’s widespread support.

And professional photographers don’t want lossy compression anywhere near their cameras.

Asa Dotzler, on July 15, 2009 at 5:23 pm, said:

I think I’m in a better position than you to speak to what Mozilla does and doesn’t care about. Anyway, as you’re now into reading and reporting on the hearts and minds of people you don’t even know, I think I’m through with responding to you.

Actions speak louder than words, and words speak louder than thoughts. And from Mozilla/Xiph’s actions and words, the only thing they’ve shown themselves to care about with regard to video codecs is patents.

Though that was a bit harsh on Mozilla; after all they did pay Xiph to get their butts in gear and upgrade Theora from pathetic last generation tech to mere last generation tech.

Google’s been pushing Chrome not just on YouTube, but with a giant and first of a kind banner ad on the front page of Google.com for months now. The banner ad has a big Chrome log, a big “Install Google Chrome” button and the message “A faster way to browse the web”. It’s had only a marginal impact on Chrome’s growth and couldn’t even lift Chrome up to Firefox’s early growth rate.

I think you wildly over-estimate the general Web population’s willingness to download, install, and use software replacements. We’ve been fighting this complacency for a lot longer than Google and it’s just not an easy task to get people to change.

“Or, to get lots of people switching to HTML5-capable browsers, you need some kickass <video> implementations to give them a reason to switch. I’m not sure we’re in a position to say which is true just yet.”

I’ve yet to see content driving browser adoption in any significant numbers anywhere since the first progressive images hit the Web. I’ve been working on this problem for a decade with Mozilla and new and awesome content just doesn’t cause rapid migration to new software clients. Even plug-ins like Flash didn’t get very far until Microsoft bundled it with Windows ensuring pretty much everyone got it. And it’s even worse because there are very few people or companies willing to invest in new and awesome content that is only viewable by small fractions of the Web population.

The Web browsing population evolves slowly. That’s just a fact. There are a lot of chicken-and-egg problems and most of the ones I’ve seen become successful were the result of a broad platform, not a superior kind of content.

So there are codecs that give 10x the compression for the same quality as today’s JPEGs? If it’s 10 times better, surely Nikon and Canon are shipping it in all their most popular cameras. Right?

“It’s just sad that Mozilla and Xiph don’t care about video technology”

I think I’m in a better position than you to speak to what Mozilla does and doesn’t care about. Anyway, as you’re now into reading and reporting on the hearts and minds of people you don’t even know, I think I’m through with responding to you.

“I don’t see large numbers of users clamoring for something better. Heck, I don’t see large numbers of content producers clamoring for something better.

Most people don’t realize that you can get 10x better than JPEG with modern codecs. And those that do also realize that also realize that the img tag implementations are frozen to JPEG, GIF, and PNG for all eternity. And on the topic of content providers, both Google and Facebook at least are very interested in saving image bandwidth. There just isn’t anything they can do but use JPEG.

As for the rise of Flash, you’re right in that it was because it was the only thing that worked everywhere and had *nothing* to do with how good the technology was. This is the exact same reason that if Theora becomes the standard for the video tag, then that’s what the video tag will be used for. And given the history of the img tag, that’s all it will be used for for the next decade and a half.

Given Thusnelda still adds quite visible artifacts even at max quality on half the videos I’ve tried in it (and that’s ignoring how badly it massacres just about any HD video I feed it), any situation that involves using Theora that long is deplorable given that vastly better codecs already exist and are already being widely adopted as the new standard for web video.

And if the video tag was really going to be a threat against Flash, it would have supported H.264 so that websites could have a painless transition from Flash instead of having to go back and transcode their whole video library and have a long period with them serving up two formats.

It’s just sad that Mozilla and Xiph don’t care about video technology and would rather keep web video as seas of blocks, rather than push forward with better technology. If Xiph had done *something* to further open video codecs other than sit on their asses for years after slapping a few unused features onto VP3 and calling it Theora, this whole situation wouldn’t be so pathetic.

It sounds like you’re arguing that the best codec trumps those other considerations. Is that a fair characterization?

Yeah, I think that’s fair. I’m well aware of the problems, I just place different emphasis on aspects of the equation. I’m also pretty confident that if h.264 becomes a common format on the web, Mozilla will find a way to support it. Perhaps because by then certain licensing restrictions would almost certainly have already had to be overcome.

Asa Dotzler, on July 15, 2009 at 1:37 am, said:

When YouTube launched, it didn’t have the best video technology. It had, arguably, some of the worst, with Flash 6′s Sorenson Spark and mono mp3 audio and at god-awful bitrates. They opted for the technology equivalent of what was shipping in RealPlayer way back in 1997 and in QuickTime in 1998.

And look what we’re stuck with. Video so poor it’s left in the dust by h.264 and Theora! Not a happy place to be.

If you really want HTML5′s <video> support to succeed, you first need lots and lots of people using a <video>-capable browser.

Maybe. Or, to get lots of people switching to HTML5-capable browsers, you need some kickass <video> implementations to give them a reason to switch. I’m not sure we’re in a position to say which is true just yet.

Firefox and Opera and Chrome combined, on the other hand, account for 360 million users and over 30% of Web usage. That’s starting to look like to enough reach to convince Web developers and content producers to move off of Flash or at least start supplementing it with HTML5 <video>.

I’m not exactly sure where your numbers are from, but Net Applications – which tends to overestimate the market share of decent browsers like FF, Opera et al – would put those combined figures at about 25%, and Chrome + Safari at 10%. Either grouping would be the underdog.

I can’t help but think people underestimate Google on this one. Their brand rather crosses boundaries, in a way that Safari (mainly Mac users) and Firefox (largely enlightened Windows users and Linux types) just can’t; and they own YouTube. If they were to push Chrome on YouTube, everybody’s mother and aunt Flossie would trust them – because they do what they’re told and Google must know best, right? It ties into what Luke says, too:

Luke Mildenhall-Ward, on July 15, 2009 at 1:44 pm, said:

If there’s going to be a war, it’ll end up badly for everyone. But if there was going to be a war, the winner would be determined by the content providers.

As he points out, Google seem to like h.264; and YouTube already has a giant library of h.264 content for compatibility with mobile devices. They’ll clearly make that available, and probably push it hard. Theora, on the other hand, involves transcoding their entire library… I kind of suspect, too, that many content providers will follow YouTube if they’re successful in whatever <video> solution they provide. People ape successful businesses, even if the success isn’t really due to the exact implementation details.

John Michael, on July 15, 2009 at 12:11 am, said:

Ironically, this is one strength of Flash – only one company controls the implementation so they can upgrade codecs without all this worry of getting a multitude of browser vendors to support it.

It’s interesting, isn’t it? Closed platforms can be great for consumers in some ways, and death in others. The real trouble seems to come when you get a successful platform that doesn’t share your priorities; like, say, “not crashing my web browser all the sodding time”, or “allowing copy and paste”…

I certainly don’t want to lose sight of the fact that whatever happens, it beats today’s norm of dubious-quality plugins. I trust that we’ll have a clear winner, or at least a clearly winning solution (which is frankly in sight, with the excellent Video for Everybody), and that all sides will pull together to evangelise it.

So you think that content producers are going to adopt the video tag and a codec that’s only available for less than 10% of the Web and that’s going to cause those sub-10% browsers to grow their share dramatically?

That sounds like a pretty unlikely future to me.

“The dominance of Flash was originally rooted by the content providers, and it proved users will download additional software if they want the content enough.”

Not really. Flash was ubiquitous because Microsoft shipped it on XP and where they didn’t ship it, OEMs added it. It was everywhere and that’s why content producers used it. Not the other way around.

“Use YouTube in higher quality, more efficiently, and much faster by downloading Chrome”, we would easily see a huge proportion of web users switch browsers overnight,”

YouTube and Google have already done this with banner ads on their sites advertising Chrome for several months now and as best I can tell it took their share of Web browser usage from ~1.5 to ~2.0%. I think that’s far from a “huge proportion of web users”.

I think you think the Web moves much faster and more dramatically than it actually does.

If there’s going to be a war, it’ll end up badly for everyone. But if there was going to be a war, the winner would be determined by the content providers.

I think it’s also important to consider the impending mobile internet boom. And we all know who’s the dominant player in that sector…

JPEG has been and is serving the Web quite well. I don’t see large numbers of users clamoring for something better. Heck, I don’t see large numbers of content producers clamoring for something better. Good enough is often good enough.

I think that the YouTube explosion of 2005 is relevant here.

When YouTube launched, it didn’t have the best video technology. It had, arguably, some of the worst, with Flash 6′s Sorenson Spark and mono mp3 audio and at god-awful bitrates. They opted for the technology equivalent of what was shipping in RealPlayer way back in 1997 and in QuickTime in 1998.

But what it didn’t have in great video quality it more than made up for with Flash’s ubiquity. YouTube videos were pretty crap, but “everyone” could view them without downloading a plug-in or waiting very long for the content to buffer.

And YouTube videos continued to be crap for several years because download wait time and player ubiquity trumped everything. And users loved it!

There were higher quality solutions available when YouTube launched. QuickTime had shipped h.264 in v.7 that spring but YouTube didn’t use it. They opted for the decade old tech, but more widely deployed solution. Microsoft had shipped WMP v.10 with the WMV9 codec almost a year before YouTube launched but they didn’t use that either. Again, they opted for the much worse and much older tech. Both Apple’s and Microsoft’s solutions were far superior to Flash’s video capability in every way but neither had the installed base of Flash. So, given a choice between the best available technology a 10 year old codec that was more widely deployed, they went with the 10 year old codec.

Sometimes good enough is actually good enough. Sometimes worse is even better.

As a baseline codec, Theora could serve the Web well. It could provide an experience that is more than sufficient for all but the most discriminating users. And for those users, no one is trying to deny them H.264 or H.265 or whatever is the next cutting edge solution.