Ananova Gives Web "Human" Face
For the past year the British news agency Press Association (PA) has been conceiving, programming, writing, and designing a new collective image of itself for the Internet. Recently, it appeared in the form of Ananova.com. Featuring a text-to-speech engine designed to read news stories while a 3D engine running in parallel animates a human-like face, Ananova.com bills its eponymous streaming-media product as the world's first virtual newscaster.
Designed to look like a hip young Brit (with green-tinted hair, no less), the new talking head admits to a fondness for "The Simpsons," Oasis, and the sport of cricket, according to the PA. But when it comes to news, she's all business. Speaking in a North-Atlantic accent, her reports cover breaking news from around the world. And she responds to everything she says.
The key to Ananova is that she is not only a breakthrough in the news industry but employs breakthrough technology to broadcast her information feed.
Ananova is rendered via a real-time system that tells her what to do, what to say, and how to say it. She has her own search engine to help her track down the latest news, according to the PA. As information arrives, it is tagged in XML to format her audiovisual cues. The tagged information is then fed to the text-to-speech application that creates her voice, while her live animation code matches her vocal movements to her words.
Her chief developer says the PA chose to use XML in an internally built publishing system because of its readability. "Nodes are stackable and nestable. With the correct DTD, you can do subsequent processing," says Jonathon Jowitz, head of project management at Ananova.com.
XML gives the project the potential to render parsed files in various formats on various devices. "We can convert them in output mode to drive the device we're talking to. QuickTime, Windows. We can do anything we want at the output stage."
Jowitz notes, though, that there are scalability constraints on the product, so files when streamed out are very small, either 8 or 19kb streams. "The biggest problem has been far too many people for the number of servers brought into action." So there's a net congestion signal you might see while viewing a Webcast. "Every time you put someone at 200k, you lose 10 at 19. It's pure math. So, we're working on getting the product better, getting the delivery better, honing the engines. These are all parallel processes. The plans for the future are virtually limitless. We want to give them what they want, as opposed to feeding them."
The news service currently presents only a handful of news items on demand, but Ananova's builders envision her becoming increasingly ubiquitous and multifunctional.
"If you're too busy to keep checking the site, I'll come to you," writes Ananova on the site. "Tell me what you're interested in and I'll send you an e-mail as soon as I find it." The first version offers a selection of e-mail updates in three general news categories.
With the emerging generation of Webtop services for content aggregation, it's easy to envision Ananova becoming a personal information portal, selecting stories based on user preferences and mixing them into a timesaving information stream.
"The Internet is the first step," says Deborah Stephens, from the PA's press office. "We envision her in phones, in wristwatches, in alarm clocks. The wireless solutions depend on the development of the technology, but we're thinking in terms of months, not years."
Using many tools, Ananova's programmers wrote her synched systems, with the help of co-partners, primarily in C++. Jowitz notes that Ananova is many nested programs wrapped in a master program.
He says he has no notion of how many lines of code it took to create her, but "millions is the brief answer."
"The phoneme library is 200 megs. The animation library is 38 megs, uncompressed. Lines of code pale in comparison to some of the pools of data that she calls upon."
"Hello world" are Ananova's first words as it introduces itself during bulletins streamed on-demand via RealPlayer 7, taking its tip for a greeting from the classic network-programming intro.
"Here is the news, and this time it's personal," she pronounces as Ananova launches into a report of a recent jet-crash. This is displaced by an image of the President of Zimbabwe in the midst of a land-reform crisis. For a brief moment, it seems like Ananova has the ability, by casting its ironic glance across the world of human affairs, to remind its viewers that, unlike its stylized self, we are fragile human beings with real lives. The product hints at something that no other news source can offer, digital irony.
The launch of Ananova (whose name could translate as "New Again" in mixed Greek and Latin) raises the question of whether Web surfers will prefer to get their news from a digital personality, as opposed to a real person.
One expert doubts they will. "I've seen other work like this, one from a company in Sweden," says John Hobbs, head of creative development at Ericsson CyberLab NY. "[It] was intended as an information agent. The idea was that by incorporating speech recognition and natural language parsing a user could query a database and the character would respond to the questions. As far as using something like this as a newsreader, I'm not convinced. An interactive responsive agent, yes, but what's the gain of having a cartoon read the news?"
Hobbs says his biggest concern is that Ananova is taking nonlinear information and converting it into linear form, drawing attention to the fact that this might be a counter-intuitive interface for content flowing through a hypertextual arena. "My personal feeling is that we should be working to go the other way. Humans are much better at absorbing non-linear data."
Another convergence specialist agrees. "No, I don't think we want to get news from digital personalities," says Dan Rayburn, manager of Streaming Media Worldwide at Globix Corp. "When users watch news on TV they do so because they can sit back and do nothing and get info. Users that go onto the Internet to get the news are looking for it and doing some work. I think those are two completely different users who are accessing news feeds, be it traditional broadcast TV or the Internet."
My sister Zainub Ashraf, associate marketing manager at NBC Internet, feels differently. "It seems pretty natural since everything seems to be going digital. At the onset, however, I don't necessarily think people are going to be 100% comfortable relying on it for their news updates. If Ananova is able to communicate accurate news briefs in real-time there is definitely a market out there. I'd tune in for sure."
The Cyborg in the Mirror
When I ask a colleague for his opinion, he wonders whether viewers will be jarred by Ananova reporting news of human suffering. From my viewpoint, as both a reporter of and consumer of news in this case, the appearance of Ananova reveals the extent to which news of human suffering has become somewhat mundane and even uninteresting. Maybe Ananova can contribute the fearless precision of a digital entity to the daily task of reflecting human suffering.
If the developers at Ananova.com are able to build an interface that maps human emotions to news, we'll be able to watch the event horizon of a stock market crash or an episode of ethnic cleansing spread across her digital face - perhaps restoring some empathy in us. If they can make her meet halfway between Christiane Amanpour and Max Headroom, I believe that Ananova stands a chance of being a valuable new service. It's up to us whether or not we like what her mirror reveals.
Technology with... |
Jonathon Jowitz, Head of Project Management
About the author:Adnan Ashraf is a producer and reporter at EarthWeb in New York City.