jd/adobe

Some early notes, after reading 'way too much all week.... ;-)

Biggest takeaway: People like rich video experiences. The big sitback screen is still first choice... broadcast served far more traffic than Web. Pundits who argue "Web vs TV" are missing that it's "Web *and* TV". But when people can experience a "Video RIA" they like it. Good validation.

But when people are excluded, they don't like it. Microsoft was heard as saying "we're bringing Olympics to the world", and only later people realized this was a US-only deal. Linux users were cut out, as were Mac/PPC owners. Then 10% of US broadband folks were cut out atop that. Microsoft would have drawn less criticism were they a little more realistic in setting expectations.

What are the numbers for Silverlight? Hard to say... still seem contradictory. Nielsen Online says in an Aug 13 press release that the video section of NBCOlympics.com received 2,030,000 unique visitors on Mon Aug 11. Microsoft is saying they got eight million "downloads" one day. When you combine geo-restriction, platform-restriction, and failed installations, the NBC site may have prompted a million successful installations one day. Looks teeny.

Whatever the actual numbers turn out to be, it doesn't seem to mean much for making Silverlight deployment to the general public any more practical... a site would still have to eat those support costs. I risk turning RIAstats.com into a gaming target by mentioning it, but Saturday morning still shows less than 2% Silverlight 2 support. The DNC doesn't seem like it will change this either. The numbers are still fuzzy, but it seems pretty clear Silverlight's silver bullet shot blanks.

Still unclear to me is the mobile angle. Some US-oriented quotes seem to show this at 25% of the desktop browser video viewing. Considering there are probably device restrictions, atop the OS restrictions and geo-restrictions, this could be a big deal. Needs more detail.

Also unclear so far is the overall global picture, and how people worldwide actually used web video this Olympics. China has a bigger internet audience than the US, and much more interest in the games themselves... news services uniformly use Flash video these days... regional licensees seemed to mostly deliver in non-beta software their audiences could actually view... there was massive peer-to-peer delivery this time as well.

It will take awhile for the world to really understand this worldwide video event. Signs look good that it changed expectations in a positive and useful way. We humans do like smarter video. Good sign.

Two other bits this week, Microsoft-related, but not Olympics:

ECMAScript fell down and went boom. The best numbers I've seen show IE6 at 40% marketshare, IE7 at 40%, and Opera/Firefox/Safari/etc at 20%. That's the real world. For the specification process, ECMAScript has been working on its next version for almost a decade. It's been clear for a year Microsoft won't implement it, and so the world won't support it. End of story. HTML, CSS and JavaScript continue to evolve relatively slowly. Makes the whole VIDEO/RIA/Aurora predictions seem even more unrealistic. [nb: I rewrote this paragraph an hour after initial post.]

ISO fell down and went boom, too. Microsoft pushed through the OOXML proposal. Doesn't matter that no one can implement it, and perhaps no one might even want to implement it... Microsoft Office is no longer barred from governmental purchase because it's not a politically-mandated "open standard". Circus all around on that one.

Put those two items together and it gets really silly... Microsoft saying "ooh ES4 is too hard for us to implement" (despite it being already deployed to over 90% of consumer machines today!), then pushing through "an open standard" that even they can't implement. Just business, not personal.

Anyway, for in-the-browser delivery, it's still "Flash Just Works". I can understand that committed .NET developers might want to believe otherwise, and those heavily invested in cross-browser JavaScript 1.x frameworks might want to believe otherwise, but no amount of bloviation changes the basics. Adobe Flash Player provides universal publishing capability, and truly rapid evolution atop that. The Adobe Integrated Runtime is bringing this beyond-the-browser, to trusted Internet apps. Flash Just Works.

And people do indeed like live video communications. The trend's our friend.

Posted by John Dowdell at 07:50 AM | Permalink | Comments (19)

If you're ever deciding between On2 VP6 and H.264, then here is info on how the BBC went about it.

I micro-blogged this earlier today on Twitter, but want to call out some main topics in the weblog.

An intro to video delivery choices:

The video you see in BBC iPlayer today is encoded using the On2 VP6 codec, at a bitrate of 500Kbps. The On2 codec (a video compression technology from a company called On2) is pretty much the standard for video delivery over the internet today. It's optimised for moderately low data rates (300Kbps to 700Kbps, rather than the 2Mbps to 4Mbps needed for HD content), and low CPU usage, allowing it to work reasonably well on older computers. In short, On2 VP6 is the video workhorse of the internet.

... Compared to On2 VP6, H.264 delivers sharper video quality at a lower data rate, but requires more CPU power to decode, particularly on older machines, and the user needs to have the latest version of Flash installed.

Back in December of last year, relatively few people had installed the Flash player needed to play H.264 content; now almost 80% of BBC iPlayer users have it. More machines now have graphics cards with H.264 hardware acceleration. Additionally, Level3, a content distribution network (CDN) is now able to stream H.264 content to ISPs in the UK, and the content encoding workflows that we use (Anystream and Telestream) are now able to support H.264.

... The good news for those looking for video quality improvements in BBC iPlayer is that, starting this week, we're going to be encoding our content in H.264 format at 800Kbps. Additionally, our media player now supports hardware acceleration in full-screen mode, giving a greatly improved image at lower CPU usage than before.

So they've got the clientside runtime technology already installed (Adobe Flash Player), and the production workflow almost migrated (changing to MainConcept encoders), and their content distribution network is about ready to go H.264 too.

Final element? User experience. You can't yank peoples' habits, expectations out from under them. That's why the release will be in stages. First stage is offering parallel VP6 and H264, with VP6 as default, and H264 available via a "Play high quality" button. Once this is realworld-tested, the next stage is to turn on automatic bitrate detection, meaning that H264 will become the default on good connections. The stage after that would be analyzing bandwidth changes and audience desire. They're getting their feedback a little at a time, not asking the viewing audience to change to too much, too quickly, without recourse.

Also see Erik Huggers, who gives the larger picture about the move.

In comments at Anthony Rose's technical discussion: "Is this new codec going to be compatable with the Nintendo Wii?" This is a tough question... but it's a valid question. iPhone and PlayStation owners ask the same thing. Nokia Internet Tablet, iRiver, and many other devices achieve standard capability via Adobe Flash Player. But it did take awhile before office printers standardized on Adobe PostScript... there will always devices which don't include standard capabilities, especially during the early days.

Innovative file-format types do tend to be commodified over time... bitmap formats work better across devices now, and text is easier than in the early years. Mozilla will be adding the On2 VP3 codec next year, as has Opera. But I imagine it would be expensive for realworld video production workflows to distribute an additional older format of compressed video for a minority audience... desirable, sure, but expensive. See how it goes.

You've got to get all four legs of the stool solid: the production workflow, the distribution process, the clientside capability, and then the user experience. The BBC is a good example of how a video production group actually goes about this testing. I'm glad the BBC is so open about how they're bringing about this work.

Posted by John Dowdell at 07:24 PM | Permalink | Comments (0)

Adobe Media Player had a new release yesterday, with performance, interface, and publishing improvements. Pundits covered the press release. Most of them focused on which shows to watch. Some spoke of "competitors like Apple's iTunes and Windows Media Player." Seems to me like they're only seeing one small part of the picture.

I'm not putting down the pundits... they're entitled to view anything anyway they see fit, and it's likely that Adobe hasn't done the strongest job in insisting on understanding of its goals.

But this isn't a story like competing cable or satellite providers, and which company secures the rights to a given title in a given region. It's not one of those earlier zero-sum games about which movie theatre chains can show the fancy MGM films, and which chains shows the quicker RKO titles. It's more like designing the first cathode-ray tubes, or discovering the business models for the first TV stations. We know we have to achieve something, but still have to figure out its final form.

Off the top of my head, Adobe Media Player 1.0 delivered three significant things: a local Rich Internet Application for viewing video from diverse sources, replacing the page-refresh video experience where your viewing history is stored on some company's server and then data-mined... a subscription-based, "video comes to you" experience, laying the groundwork for future decentralized social viewing recommendations... and a policy-file system which lets content creators determine where they grant viewing rights, to whom, and when.

There's a whole lot more technology understructure yet to test and develop. We need better creation, integration, and analysis of video metadata, a coherent pipeline for including information along with the visuals. More work needs to be done on the overall production workflow, from planning to asset tracking to editing to compression to distribution to analytics. We need to analyze and expose video content (speech-to-text, object recognition) and social metadata (recommendations and tagging from friends), so that advertising (the main revenue generator) can be more tuned to the viewing audience's current needs. We need to make the video experience more customizable to the content producer's experiential design, and more responsive to the enduser's desires. AMP 1.x did accomplish a lot, but there's much more work yet to done.

Within Adobe I don't hear people talk of competing with particular video sites and distributors... such initiatives are uniformly seen as partners, or as potential partners, who have technology needs which are not yet met, perhaps not yet even recognized.

Everything I see points towards Adobe Media Player as being a technology platform, where new publishing capabilities can be explored and refined. "How will the viewing public want to experience video in 2010, 2015, 2020?" seems more the spirit than "What's the marketshare for the catalog this month?" That's the reality I see within Adobe, from the product team to the Dynamic Media organization to the top executive level. We need to bring about the next generation of internet video publishing.

Back when Adobe PostScript first provided a more predictable digital printing capability than the maze of competing printer drivers and application exporters, Adobe Systems did not open a publishing house or try to monopolize book distribution. Adobe achieved its financial goals by opening up the publishing process to universal access, earning money on sales of enabling technology.

To confirm this cultural DNA, see this recent interview with Charles Geschke, co-founder of Adobe. He was a math professor who found computers personally interesting, found great satisfaction in making the written word more accessible to more people than anyone since Gutenberg, and now stresses the importance of the new integrated runtime for networked application development. It's a longterm view, not about controlling content, but about profiting by helping others sustainably create and use content. This cultural orientation was one of the most striking things I learned about Adobe, coming in from the cowboy/interactivity/shrinkwrap culture at Macromedia.

Other companies have different business models, whether it's to sell a range of consumer electronic devices, or to develop an advertising network, or to increase the range of proprietary personalization databases. But Adobe makes its money selling neutral publishing technology. The more people who find video useful, the better. These core orientations determine the different paths each group will take.

The AMP marketing materials do emphasize significant partners, but there seem to be two main reasons for that: (a) we need a goodsized audience to provide reliable feedback on how people will really want things to work; and (b) creators are actively seeking ways to maintain a connection with their big-budget work, beyond just letting an .MPG run wild throughout the Web in hope that money will somehow flow in. Sony, CBS, Showtime, MTV and all the rest are validating that they find potential within AMP, and those big titles help draw useful audiences during development.

(There's an interesting angle on big-budget video... YouTube became famous during the Web 2.0 age for "User Generated Content" (UGC). That's great, I love it, it's valuable, we need it. But audiences tend to obsess on professional, big-budget entertainment... that's they stuff they hunt for. And to afford those professional budgets, creators need a real return-on-investment. Finding ways to support video-creation businesses is the real way to support video-viewing audiences.)

Anyway, when I see the daily commentary focus on Ghostbusters or The Love Boat, I can understand how someone fresh to it may just see what's personally relevant for them. But I see bigger goals within Adobe, focused far more on new technology than commercial properties.

The Adobe video pipeline is undergoing startling rapid evolution. I'm betting the Techmeme stories from Summer 2009 will likely be quite a bit different.... ;-)

Posted by John Dowdell at 11:46 AM | Permalink | Comments (0)

Max Bloom has a great article at StreamingMedia.com, discussing aspects of video delivery for the Beijing Olympics this summer. I've been tracking this project, because of all the prior publicity about a "Flash Killer". I found this article after reading of Akamai's delivery of Olympics video (more), and the StreamingMedia.com piece has lots of detail I hadn't seen before.

Main takeaway? Like MLB.com, Silverlight is offered as a new option to an existing WMV9 workflow, and Windows Media Player will still work as a client.

(The MLB.com deal achieved a lot of positive publicity for Microsoft in 2007, but the reception upon release was not as positive, and today MLB.com doesn't even seem to appear in Microsoft marketing materials. All through this, MLB Gameday has continued innovating in Flash.)

Regional restrictions enable regional licensing, and so enable funding for the event itself. A YouTube-style model wouldn't work for such a large, capital-intensive event. Adobe Media Player offers similar capabilities today, but the WMV architecture has been in trials for the past few Olympics. The entire deal makes more sense when seen in this light.

Other details include the hiring of Schematic and the business angles for UI design, how viewing varies across different regions, and this bit on the details of delivery: "As of press time, NBCOlympics.com had yet to make a number of important technical decisions. A slew of DRC-Stream software and encoder boards from Canada-based Digital Rapids are being deployed in Beijing to populate NBCOlympics.com’s encoding farm, but other than committing to VC-1, NBCOlympics.com has yet to confirm encoding bitrates, frame rates, or frame sizes. (Without offering more specifics, Miller says NBCOlympics.com will be streaming through a managed bitrate solution to optimize the user’s connection, with a target maximum bitrate of 650KB/sec.) Digital Rapids is also supplying software to enable transcoding from other digital media formats into VC-1."

After reading the article and sleeping on it, I'm left with the impression that this is less "Microsoft buying audience share by subsidizing big events" than it is "existing Windows Media sites not defecting because there's now an option for in-the-page viewing". I'm guessing Silverlight will still receive a boost from the Olympics deal, but it doesn't seem as dramatic as it did before the details became available.

Posted by John Dowdell at 10:32 AM | Permalink | Comments (1)