http://www.wbez.org/Content.aspx?audioID=14285

While this story talks about news reporting, you just gotta know that somewhere, somebody's working on this for a radio application...

then again, who needs people at all?

People who need people are the luckiest people in the world...

Love,

Barbra

(You knew it was coming...)

***WARNING! CRAPPY 80'S REFERENCE AHEAD!!!***

That anchorperson in the photo could be Max Headroom's cousin.

What, No thanks for the Wonderful set up?!?!

Seems a tad ungrateful!

STOP IT! Quit giving CHEAP CHANNEL MORE IDEAS to cheapen and degrade their product.

I guess that this means that fans of ¢harlie FM aren't amongst "the luckiest people in the world."

heehee.....Max would be proud...............

It's good, yeah, but it's severely lacking appeal due to the absence of the "laser lights" background. A plain brown backdrop seems uninviting.

Do ya think this thing generates "ZikZak" ads as well? You know, the kind that make people explode when they see them???

Seriously tho: in radio (where an animated, automated face image wouldn't be fe-fe-feasible) the concept of using computer-generated on-air "talent" is completely do-able. You'd type up the text to read (liners, spots etc.) and the computer would read them on-air timed by pre-programmed cues. It's called "text to speech."

Ultimately computer speech (in general) will progress to the point where it would be indistinguishable from a human voice. It's not there yet, but they are working on it. There are packages which do use actual human speech samples and play them back to form speech, in much the same way a wavetable synthesiser uses actual instrument samples to create music. (An entire orchestra on your desktop. Imagine that!!) Speechify is one, Socrates "Talking System" is another one--albeit somewhat ancient--that I know of. Trouble with those systems is because they are sample-based, they still sound somewhat "robotic."

But just think of it like this: NOAA's been broadcasting this way for several years now..............

Now, *why* anybody would actually want to apply this to mainstream broadcast radio outside of the NOAA camp is definately open for question, unless broadcasters (read: Cheap Channel) want to make radio listening even crappier than it already is.

Just a thought.........

On a more serious note, I think that the public's embracing of cartoons and computer generated graphics in entertainment (i.e. the Shrek movies, Fox's Sunday night lineup, commercials in which images of actors have been converted into cartoons by computers, etc.) is setting the stage for something like this. Perhaps, this technology will first be applied to home-shopping broadcasts before it becomes acceptable to use it for reporting news.

In a very loose sense, the science fiction writers of the 1950s were right when they predicted that one day people would listen to and interact with "robots." One thing that they missed was that many of these "robots" would be virtual, rather than mechanical androids. Another thing that the futurists generally overlooked was what happened to the people who were formerly doing the work that was taken over by machines.

Well take a look at the latest CGI movie due out. http://www.beowulfmovie.com/ Emotions will never be computer generated with very good accuracy and when a human makes a humorous mistake on air, well that is just good radio. The funny goof's will never take place if there is just a robot driving the silly thing. My Opinion of course.

Wouldn't you still need people to tell it what to say? And if that's the case, why not just have that person say it?

Exactly!

A year or so ago, I was reading about software companies that were selling talking cartoon characters as customer service agents for your Web site. They were really serious.

I don't know what it is like in the U.S., but the Canadian futureshop.ca (they own Best Buy in Canada, too, and have most of the same products) has a a very cool Black dude who wants to answer any of your technology questions about their products. It is created by filming a real person.

I think I may have mentioned it on this forum previously, but I really do like Monkee Mail: http://www.careerbuilder.com/monk-e-mail/

An important topic regarding computer-generated people such as those used in Beowulf (or my favourite vomit-inducing flick, "The Polar Express") is the "Uncanny Valley." This is a theoretical state of mind (paradigm, in a way) wherein the more realistic attributes that computer-generated personalities assume, such as by robots or those which appear on screen, basically the more revolting they are perceived to be.

The concept generally applies to robotics but it can easily be applied to other media, such as TV shows or films. Or, closer to home, automated news broadcasting.

Needless to say, when I saw "Polar Express" a few years ago the characters really creeped me out. (And I saw it at the OMNIMAX to boot.)

Fortunately Max Headroom doesn't have the Valley problem too much (I think), as the title character himself is actually a guy in costume and makeup. Computer technology and animation in the mid-80s wasn't yet at the point where entire characters (and entire films!) could be created like today.

http://en.wikipedia.org/wiki/Uncanny_valley

"This area of repulsive response aroused by a robot with appearance and motion between a "barely-human" and "fully human" entity is called the Uncanny Valley. The name captures the idea that a robot which is "almost human" will seem overly "strange" to a human being and thus will fail to evoke the empathetic response required for productive human-robot interaction."

This, I guess, would explain why I get quite a kick out of C-3PO but look at some of the human-representative characters in some modern video games as being overly strange, especially if said games require human input i.e. via microphone.


A point not 100% related but still definately worth mentioning in this context~
Alan Alda did an episode for his "Scientific American Frontiers" TV programme, titled "The Intimate Machine", episode 1303, in which he explained how researchers at one university (MIT, I think) created interactive software which displayed the likeness of a child on screen (his name was "SAM"), with which another (human) child would interact. The computer generated child would react to the things the human child would do, creating the impression that the computer generated child was somehow human himself. To my understanding, the researchers and programmers did this as a "conceptual demonstration" of artificial intellegence. (The episode is re-run occasionally on OPB/OPAN Oregon Channel, if you have access to an ATSC receiver.)

Informationz:
http://www.pbs.org/saf/1303/segments/1303-1.htm (deals with the particular segment I described)
http://www.pbs.org/saf/1303/ (stuff dealing with the entire episode)
http://www.pbs.org/perl/media.cgir?t=w&f=virage/scientific/pbssaf1303_220k.asf (view in Winamp--NOTE you need the Nullsoft Netshow plug-in, "in_nsds.dll", in order to view it.)

It's really fascinating that technology can be applied this way (I think) but to be perfectly honest, I feel great care should be taken by those who produce these technologies so as not to make them appear too "human-like", for sake of positive human perception. The "gap between man and machine" can really only be bridged to a certain point where said bridge becomes structurally unsound, so to speak.

HAL 9000 in the morning . . .

I'm sorry Skeptical, I'm afraid I can't do that.............

*big grin*

Hal Raymond in the morning! Oops, wrong thread.

Interesting choice in that HAL killed all of his crew except for one! Is there any Irony in that?