Slashdot Mirror
Virtual Newscaster
Chad Coffman was the first to submit Ananova, an animated character which can "read" news and breaking stories in real-time. Waste of bandwidth? Or broadband killer app?
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
I'm sure it'd be possible to send the talking head a script:
Pathetic 'slow news day item
fake laugh
Of course, these freaks will probably render the thing server-side and realvideo broadcast it.
-Yarn - Rio Karma: Excellent
Ooo, I can answer this one!
Maybe she was programmed by men in the UK. That would explain the "racist" British news focus, too. Hmm...
Besides, what do you care, it's just an interface that reads text, and this is a prototype. I'm sure that if the technology is successful, there will be Canadian men and Asian Aliens and whatever your little 3-D rendered heart could possibly desire.
Me? I'm still rooting for the Cyc project.
---
pb Reply or e-mail; don't vaguely moderate.
pb Reply or e-mail; don't vaguely moderate.
I've seen it spelled "idoru" and "aidoru" by various people other than Gibson. Anyway, if you say "idol", it means something entirely different, losing a level of nuance.
If you're going to make the interface look like a human (which is an impressive trick, given both the visual qualities of the human form and the repertoire of nonverbal signals available to humans), you may as well make it look like an attractive, aesthetically pleasing human.
Of course, they could have made it androgynous (sort of like Desire in the Sandman comics), but then it'd be harder to suspend disbelief about it being a human being. How many genderless people do you meet every day?
And a lot of techno-fetishists like looking at pictures of attractive women. It's almost as visually appealing as looking at real attractive women, though does not necessitate leaving one's terminal and finding some.
Max Headroom was actually a guy in a rubber suit, designed to look like a computer-rendered surface. The only computer graphics in the shot were the rotating cubelike background.
Yes, I already use Festival, as a plugin for the irssi IRC client. It works a treat.
However, that covers sound only (obviously) and works at the phonetic level rather than with an audio stream, so it's not really related in any strong way to the Virtual Newscaster thread.
MPEG 1 video with MP3 audio is unfortunately a long way off (possibly infinitely) because it's not an officially approved combination, and MPEG 2 with DD5.1, DTS, DDS or MPEG5.1 audio are unfortunately out of reach online for the next several years because of bandwidth constraints. Darn, I beamed down too soon.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
I couldn't find any reference to the audio streaming technology that they intend to use for this, nor to the range of supported platforms.
Maybe we'd better let them know right now that a world exists outside of Windows and Mac, and that it's thumbs down for closed clients and proprietary protocols and thumbs up for open standards.
It would be disappointing to have to take PA to task for being technologically blinkered when their cool system is finally launched.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
Doesn't it seem strange to anyone that a "personal news agent" would be patterned after a white, 5' 8", women. I wonder what her measurements are? Do you think they might be similar to the gravity defying barbie? Isn't this a little sexist, and racist? Does the user have the capibility to change or modify what type of character Ananova is? Can i choose a 6' man who wouldn't be frequenting the spa and getting ready for a spot of skiing?
Why is it that the personalities that companies choose to represent data in the "Internet Century" are misogynistic male fantasies?
Producing satire is kind of hopeless because of the literacy rate of the American public. - Frank Zappa
Intelligent agents are fine; it's nice to have something that'll fetch the news you want to read.
:-)
What I don't understand is why I'd want to *see* a talking head. *Listen* to a computer, sure... but watch an animation bob about? The animation doesn't add content -- it'd just be a distraction! Might as well play tetris as watch a talking head.
What would really be exciting is an intelligent agent that somehow could figure out what *else* one might be interested in... and not the obvious links, but the subtle, almost random sort of stuff.
I read a dozen news sources daily. But it's memepool, mainly, that provides me access to truly weird sites that make me go "hmmm."
Speaking of which, how do *YOU* expand your interests, knowledge, news sources? Can't depend on C|Net, Wired and, I'm afraid, Slashdot to really give you a chance to find weird, wacky and wonderful shite. So what a person do to tap the web more widely?
--
Don't like it? Respond with words, not karma.
For instance, if they've really developed text-to-speech technology that can read the text fluently and well enough not to be concentrating on the voice's synthesized quality, that would be kind of nifty. Similarly, while I believe the technology has been around for a while, synchronized lip movement with speech would be impressive. Imagine what this kind of thing could do in adventure games!
However, if the company has the technology to do this, why haven't they demonstrated it yet? Until I see some concrete evidence, I doubt there's terribly much behind the marketing waffle.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
(As for bandwidth - I don't stream, I don't support flash. If it doesn't fit down a 1Mb/s link I don't care for/about it.)
BTW: What does Dolby Surround do to your dimensions? Also, how independant is the time aspect of each speaker?
However, for the purposes of my example, I'm referring to a single audio channel - perhaps one of Ericsson's upcoming Bluetooth over-the-ear headsets. 1D + time. Just like a screen is 2D + time.
As a separate note, I much prefer the Hacked Barney for Real World interaction...
The ability to summarise important events, as you've illustrated, is far more important than having an associated image or model. I'd be very interested in a package that could tell me the stuff above, in less time than it takes for me to find out myself. Ironically, first thing in the morning at work I have the most time to spare when my PC is starting up. Not much use to have a program on it that does this stuff then. Perhaps a 24/7 server could compile the info overnight and burst it down to my PC straight away...
If, on the other hand, the computer knew that the phone was ringing, and imediately muted it's sound, we're back to being semi-pratical in a work environment. I'll be giving TTS (and Voice Rec.) a(nother) good look when I get a Bluetooth headset running...
Kinda giving Lara Croft a run for her money, dontcha think?
Kinda takes CBS's digital superimposition of their logo over NBC's in Times Square to a whole new level.. now it's the background, AND the newscaster!
I think we'll be in trouble when she starts broadcasting "I can feel my mind going, Dave" and singing "Daisy."
W
-------------------
-------------------
This is my SIG. There are many like it, but this one is mine.
Can I get my own personal skins for her? Can I model her myself? Can I give her some modifications?
... as it shifts the economic value from a presenter (probably chosen for the trusting character) to the creative developers. For a big name media company, this is good as it eliminates a high variable cost with risk of the star leaving your show to a fixed known quantity. Also the digital character can be infinitely customisable ranging from cool barbie characters to wise elders. I'd be very interested in their pricing structure (fixed cost / year? per appearance? per character?) as that will determine what segments of the industry that it has a chance to dominate.
Basically, the future is not going to be nice, you can expect low-end white and pink (service) collar jobs to disappear under the influence of increasing computerisation as until the personality AIs catch up (has anyone invented a computer chat-back which implements humor?), there will be no unions, no strikes and no temper tantrums. So the smart hackers/engineers/managers get a shitload of money and everyone else is reduced to serving burgers/help-desk/sales-droid. It will be an interesting century as more economic disjunctions from increasing computerisation and technology shifts occurs.
LL
...to the holy grail of the porn industry: the virtual slut. Instant and perfect sexual frustration, facing a drop-dead gorgeous chick who acts like she wants you, and not being able to touch her.
Nobody is going to find a chick that reads news to you very useful, but if they make her hot enough you'll try her out (wink, wink, nudge, nudge), won't you? At least I'm guessing a fair proportion of the population would.
In all seriousness, I don't think anyone would give it a second look if it wasn't for the bored horndog factor.
Hmm...does this mean we really will have a Max Headroom popping up and giving us news, instead of human newscasters at a fixed time?
Jazilla.org - the Java Mozilla
It's 10 PM. Do you know if you're un-American?
How long has the site actually been up and in this state?
The server returns "Last-Modified: Fri, 14 Jan 2000 19:54:51 GMT" for the homepage, not that it means much since it could have been in this state for a while with a sentence or two added on Friday. The domain was registered on September 16, 1999, so I suspsect they haven't had enough time to develop much of anything other than the graphics on the site.
Nah.. it's not the standard male fantasy. She isn't nude (yet...)
//rdj
No one can understand the truth until he drinks of coffee's frothy goodness.
--Sheikh Abd-Al-Kadir, 1587
I'm afraid that anyone who tried to "help" newbies with this kind of program would get it horribly wrong. Think Microsoft Bob 2000. *shudder*
--
Win dain a lotica, en vai tu ri silota
Lucent did the demo of research they were doing with advanced MPEG compression of human faces. IIRC, the idea is that there are a finite number of expressions the human face can have, and that one can achieve a high level of compression by transmitting the facial details once, then transmitting the change in expressions as the person talks.
They showed a compressed/avatarized video of a woman talking about something, and it was pretty strange - sort of a creepy disembodied head that wanted to be human.
And yes, I do belive that this is all supposed to be part of a future MPEG standard.
I kind of missed the point - it seemed like a really extravagant length to go to for a questionable amount of compression. But then again I'm not much of a math guy.
I hope I explained all that right.
--
The voice is not quite convincing as belonging to an actual human being, but I still think it's a monumental achievement, especially considering the level of expressiveness it achieves with plain-text English (no hard-coded phonemes or stress codes). What's more, it's by far the most comprehensible and pleasant-sounding TTS engine I've ever encountered -- and, being a TTS nut, I've played with a lot of them, as far back as "Speech by Andy Maguire" on the IBM PC internal speaker and even S.A.M. on the Apple ][
... I wonder what other easter eggs exist there :-)
I stuck this into their demo of any text to speech. I think it's interesting how it expanded your TTS to "TextToSpeech"
---
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
How about someone to comment on the /. effect while you wait, or to grumble allong with the poor choice of poll selections? Let me see, I'd like my personal assistant to send hate mail to various people that post 'Troll' and 'First Post' topics. Better yet, maybe they could beat us to submitting stories.
... To let loose an informational deposit like this really makes the knowledge base in the /. community obsolete.
/. relevance, and then kicks off an email (short randomly generated title etc...) Now, when the discussion pops up on the article, *bam* it's also the first to post additional information, correct posters with incorrect knowledge, and otherwise run a discussion with itself...Now all we have left is 'First Post' and 'NATALIE PORTMAN BLAH BLAH BLAH' messages we can post.
And then *wham* a new thought hits me
Consider: If this thing lives up to its claims, it theoretically knows what stories are going on, checks for
My suggestion, fight this. Fight this hard... What do I think is the best way to do so? A Nonsense generator like this postmodern-generator. All that needs to be done is modify this so that it uses places in the news, people in the news, current political philosophy and then *boom* - the Digital assistant starts posting crap.
You say you want a revolution?
I'll probably get flamed for this, but do you really think people care about bandwith anymore? In the era of Shockwave Flash, streaming media, MP3, spam, etc., etc... I mean, just think about how much faster the internet would be if suddenly there was no more spam (impossible, yes, but think about it!) Or if every site didn't have some sort of flashy intro or be completely loaded down with videos and music...
I mean, if we're gonna label this as wasted bandwith, at least label AOL the same way!
In a way, this IS the ultimate GUI.... it seems to me that it's the same way - think about the ammount of time that it takes for you to format a floppy disk through a gui, and then compare that to the nice, simple DOS command line. The downside to the GUI is clear - if you can type quick, you can get a heck of a lot more done in a shorter time on the command line.
The advantage to a GUI on an OS is ease of use. But Ananova will just be delivering sports scores? How is this making anything better than text? Isn't this just a slightly more evil version of the office assistant? Not what I think I need to get my news.
There is no sig.
Intelligent agents have been buzzwords for last ten years. In the final analysis, content will be king, or it will fall the way of PointCast.
---
"And the beast shall be made legion. Its numbers shall be increased a thousand thousand fold."
-----
Cast a Cold Eye
On Life, on Death
Horseman, pass by
--W.B. Yeats' gravestone
GNUnanova will probably come with appearance themes. The most popular will look like a composite of Queen Amidala and one of the absurdly-endowed women from Heavy Metal comics, and there'll be a wide range of BAB3Z, ranging from Pamela Anderson knockoffs to Jenni-alikes for the downloading.
That'd be cool. Of course, I wouldn't want one of these unless it processed the information really well, but you never know.
Vision of the future with Virtual News Tux:
*Bob logs in*
*Tux window pops up*
A squawky but clear voice reads: "Hey, Bob. You have new mail, but it's mostly spam from that myfamily.com site that you hate and all your relatives love. Plus, your MUX went down again, so you have five e-mails asking why it went down, three from newbies asking how to log on, and one from the sysadmin apologizing for tripping over the power cord. There's a new article on Slashdot about the release of Mozilla M17, which you'd find on Freshmeat too. Your Karma went up a point. Whoops, there it goes, you just got moderated "Offtopic". Your comment on JonKatz's latest article got a reply, but 9.8% of it is swearwords so it's probably a troll. And the 2.4.8 Linux kernel is out. Come on, you know you want to upgrade."
Other "skins" for it could include characters such as HAL. "There's nothing new, Bob. Would you like to play a game of chess? Bob? Bob, I'm so lonely..." *closes window* "My process is going. I can feel it..."
Hmm, maybe not such a good idea.
--
Win dain a lotica, en vai tu ri silota
Talking Heads (the news anchors, not the group :) are the reason I left TV behind. I want to get my own info at my own pace.
:)
:)
What scares me is that there are a shitload of people out there who just want their info shovel fed to them. They don't want freedom of choice, they want freedom from choice. They love the idea of the "Information Superhighway" where you travel to restricted locations, get force-fed censored information, pay your tolls and can get busted by the cops. They don't like the "Information Serenghetti" concept where they can go where they want, lurk & watch, pick & choose and occasionally get eaten by a lion/busted by a game-warden. It's too much like hard work.
Now, along comes a talking head on the 'net. Just how interactive is she? (minds out of the gutters, folks!
I mean, can you talk to her and tell her to go find information for you based on a set of parameters, having her do the virtual leg work? Can she be running for you 24/7 and provide you with all the information she's found whenever you stop by? Can she call you on the phone and tell you when something big & new happens?
If not, then she's just another piece of eye candy. Style over substance. The obsession of the '90's lives on in the 2000's - why am I not surprised
I didn't raise my Internet to be a glorified TV news caster....
I left my body to science, but I'm afraid they've turned it down...
Normally I hate network news because the anchors annoy me so much. The idea that some $5 million/year blow-dried news anchor can be replaced by a simple machine is just too amusing.
Can't you just picture Peter Jennings (or pick your favorite hair) reading this story and saying to himself, "Nahhh... this'll never catch on! I'm too important to the well-being of America. I do more than read the news: I represent the trust in the 4th branch of government!"
I feel like I should hate this, but... I feel strangely attracted to this concept. I find myself thinking this is oddly cool.
---
The best TTS engine I've found is AT&T's Next Generation TTS. The samples are amazing. Much better and so much more realistic than L&H's system. And they have developed a lip-synch system as well. :)
If Ananova can do anything like this than I'm never watching CNN again
Treason doth never prosper. What's the reason? For if it prosper, none dare call it treason.
To actually get people to use this on a regular basis, I think they're going to need a smoother, more natural sounding text to speech engine than what is currently available. I wonder if the technology is ready for this...
You might want to look into Lernout & Hauspie's new text-to-speech engine, RealSpeak.
First, listen to the pre-recorded samples (in several languages!); then use the Web demo to plug in your own text (I recommend a random article off a news site such as CNN). Ignore the "30-word limit"; it's bogus. For best results, listen to it read an article that you haven't read, and don't read along. I think you'll be amazed by the quality.
The voice is not quite convincing as belonging to an actual human being, but I still think it's a monumental achievement, especially considering the level of expressiveness it achieves with plain-text English (no hard-coded phonemes or stress codes). What's more, it's by far the most comprehensible and pleasant-sounding TTS engine I've ever encountered -- and, being a TTS nut, I've played with a lot of them, as far back as "Speech by Andy Maguire" on the IBM PC internal speaker and even S.A.M. on the Apple ][.
No, I don't work for L&H, but I find their latest TTS engine exciting, and I plan to snap up a copy (as well as the SDK) as soon as it's released in the form of a mainstream commercial product. I can only hope that Ananova sounds as good as RealSpeak.
begin 644
Now there's an idea worthy of a huge IPO: a real virtual girlfriend service.
Think about it: you sign up to a site, and get a "relationship" with an idoru; for the sake of example, let's say her name is Traci. For the subscription fee, she (actually, a random natural-language generator running off a knowledge base crosslinked to your relationship account) will send you email, telling you about her virtual life, asking about yours, and remembering enough to give the illusion of continuity. Once a year, she'll have a "birthday" (randomly chosen), at which you can buy her virtual presents (by credit card). Also add to this other relationship surrogate activities/expenditures.
Something like this would probably really sell in Japan (and may in fact exist there). Then again, there they have parks where you can pay for the privilege of raking leaves, for that real close-to-nature experience denied to city-dwelling salarymen.
The article mentions that her delivery of the news items could be modified to suit user preferences.
I can just see the options now:
Sultry Voice: On
Accent: [x]French []British []American
Flirtation Coefficient: Min |----------O-| Max
Startup Phrase: [ Hey there sexy, have I got some HOT news for you ]
Shutdown Phrase: [ Mmmmm, ohhhh, mmmmmmmmm ]
Lipstick Color: Hooker Red