The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
Presenting Kima! Your AI popstar...
WMain00
Registered User regular
Registered User regular
Reading Idoru got me thinking about the possibilities of the future and music. I'm not particularly a fan of music of today, mainly because most of it seems to be synthesized nonsense. Going through itunes the top of the music video charts in the UK is Lady GaGa recent video, a woman who in my mind is the equivalent of a blow up doll; absolutely no mind whatsoever. A product of a music industry with set ideas of what sells and what doesn't. Oh indeed GaGa will sell lots of money.
In William Gibson's Idoru, Gibson tells us of a synthetic personality known as Rei Toei; effectively an artifical intelligence designed to simulate a human female, which adapts and learns via interaction with humans. Something as complex as that is quite amazing to read considering to book was made in 1997, just as the Internet was kick starting.
Yet how far are we really from creating such a thing ourself? Imagine the possibilities. No longer would you have pay these singers, or pay for there habits and vices that follow them. Suddenly you can tell them what to do exactly as you want it to be via coding and synthesis. You could create the popstar to be whatever you want to be, and have it on live shows anywhere in the world simultaeneously.
An example of such an idea of this that never really came into fruition was Ananova, the news presenter that back when we were immature about how the Net worked we theorised was going to take over news broadcasters jobs because it would cheaper and easier to build. Of course the truth back then was that we didn't have the technology to do such a thing, but nowadays we do! What's more is we have the video community technology to publish it rapidly. To get it known and get going. Suddenly your pop star is out of a job because Kima! (random name) has come into town with new popular music and vocal chords that can range above and beyond any other pop star.
So, how far do you think we are from creating the Rei Toei's of the music world? Is it just one step away? A simple matter of fine tuning our AI technology? Or is it a storm in a teacup idea?
Reference Link: http://en.wikipedia.org/wiki/Idoru
In William Gibson's Idoru, Gibson tells us of a synthetic personality known as Rei Toei; effectively an artifical intelligence designed to simulate a human female, which adapts and learns via interaction with humans. Something as complex as that is quite amazing to read considering to book was made in 1997, just as the Internet was kick starting.
Yet how far are we really from creating such a thing ourself? Imagine the possibilities. No longer would you have pay these singers, or pay for there habits and vices that follow them. Suddenly you can tell them what to do exactly as you want it to be via coding and synthesis. You could create the popstar to be whatever you want to be, and have it on live shows anywhere in the world simultaeneously.
An example of such an idea of this that never really came into fruition was Ananova, the news presenter that back when we were immature about how the Net worked we theorised was going to take over news broadcasters jobs because it would cheaper and easier to build. Of course the truth back then was that we didn't have the technology to do such a thing, but nowadays we do! What's more is we have the video community technology to publish it rapidly. To get it known and get going. Suddenly your pop star is out of a job because Kima! (random name) has come into town with new popular music and vocal chords that can range above and beyond any other pop star.
So, how far do you think we are from creating the Rei Toei's of the music world? Is it just one step away? A simple matter of fine tuning our AI technology? Or is it a storm in a teacup idea?
Reference Link: http://en.wikipedia.org/wiki/Idoru
WMain00 on
0
Posts
ALL OF THEM, DAMMIT! This is an emergency! We were killing people just fine before the digital processor, we can do it again! My God, what do you mean the Russians are still using gauges and vacuum tubes in their aircraft?! Those clever bastards! We must close the punch-card gap!
[/totally irrelevant diatribe]
Anyway, it is quite a coincidence (given that Macross Plus came out in 1994), all the same, it's quite interesting. A case of a new, and potentially very profitable marriage of capitalism and technology, replacing capitalism's negotiation with 'art', I guess we'd call it.
I would also like to pose that good music as opposed to music that is just popular (and on rare occasions the two can overlap) has something going for it that a machine can not possess: soul. A good song makes you feel emotions, a robot cannot do that and - unless the movie AI becomes reality - never will.
http://www.youtube.com/watch?v=r4PMHt7vSE8
One of Gibson's most interesting concepts was that of machines creating art (in music or visual form) which had access to emotions that humans could not replicate. The machines evoked loneliness and sadness in ways that transcended the human experience and allowed people to feel the ache of eternities.
Honestly, what the hell is good music anyway? I suppose it is broken down into a science (certain sounds are more pleasing to the human ear than others), but music based on those sciences is not going to be universally beloved anyway.
I spend most of my childhood smashing my hands against a wood keyboard playing the 'classics'--music that, given the choice, I probably wouldn't listen to 90% of the time. So what is 'good'? Everyone has an opinion, but it's going to be subjective, when you get right down to it.
I certainly wouldn't argue that a lot of popular music is 'bad' in my opinion, but I can guarantee there's a song I think is good that other people think is bad, and vice versa. It's too subjective. Back to the computers trying to imitate human behavior....
Pop music still has soul, I think. The reason it's pop music is is able to capture commonly felt emotions or experiences, and uses that to sell the song. That's why we have all of these "yay! love!" "boo! love!" "Yay!sex" and "oh noes someone died!" songs.
http://blog.wired.com/music/2008/05/survey-produced.html
:?:
In all seriousness, it's not the "intelligence" bit that's holding back having a nonhuman pop star, it's the inability to play live shows and the CG being still slightly too bad to pass as real. It's not as though the industry is looking to create songs written by an AI, they're looking for a decent profit margin.
As for whether an AI could make decent music, of course one could. Unless your definition of AI involves something other than it being sentient.
Thanks. I'm one of those people who gets all up-in-arms when people use the word "good music". It's the product of being surrounded by people who are adamant that one form of music is 'good', then turning around and being surrounded by another group (or the same group) who are adamant that it sucks, and using the same arguments both times.
I'll accept that the emotional aspect of music is usually beneficial to the experience, so long as we aren't narrowing it down to a specific type of music (this is precisely why I almost never share what specific songs I like). Then again, you'd be hard pressed to find a song that didn't have some sort of emotional factor in it. I love the Alexandrov Red Army Choir, and that's chock full of emotion (if not the normal kind in music).
On the other hand, I could see AI having difficulty replicating that emotion sporadically (pre-rehearsed is another story). And that would be a problem.
With an AI people will know that it is totally artificial. Without at least some glimmer of the genuine, I doubt people would buy it.
Maybe the record company could lie and say it is a real person.
but they're listening to every word I say
Basically, this is it. They wouldn't want to make a pop singer without the "vices" because it's those vices that make the singer interesting to audiences. If some pop singer were to stay sober and polite throughout their career they would have a short career. Britney Spears, for example, is way past her expiration date as a musician, but she still makes people money because she seems like such a train wreck. Same for Madonna, who only stays with us because lol Kabbalah and lol adopting Africans.
What about Justin Timberlake?
but they're listening to every word I say
Cute kid, sexy videoclips, good producer, works together with artists who are also popular and good looking.
And he's a comeback kid: after nsync we didn't really expect to hear anything from them ever again, but like Robby Williams he came back.
My point is, he is not a dick. He is sober and polite throughout his career. Hell, he is on nickelodeon a whole bunch for award shows. He has had a long career.
but they're listening to every word I say
Didn't he create the Nipple-gate, though?
And yea, I don't think Cervetus' post is like a set-in-stone rule, there are many other reasons why folks can remain popular for extended periods of time.
You will hit full sentience way before you get something that can create art that humans will like.
Just to give you some context about which would happen first in AI development. Art is the final final final frontier.
Rock Band DLC | GW:OttW - arrcd | WLD - Thortar
Turns out it's all about artificial pop stars and manufactured music with no soul.
Wait...
Great soundtrack too, I have both CDs!
Was just about to bring this up. Miku (and the other voice synthesizers in the 'vocaloid' series) is popular as all hell in Japan without even any real AI. Why create a fully intelligent, sentient musical robot when you can let user-created content reign? Just add a pretty face to a voice synthesizer and Voila:
http://www.youtube.com/watch?v=QMZNKeb61Gk
You've accepted the eventuality of heuristic artificial intelligences, and you're worried about the music industry?
Please explain which of these wouldn't be possible with an AI popstar.
Cute and sexy. When was the last time you got the hots for a robot?
Robot, or digitally-generated image? After all, a physical presence isn't required. The thing doesn't have to lift boxes, after all.
And the way things are going with digital production, I wouldn't surprised if this is more than feasible in a few years.
Today.
It would be limited to pop music though. Now, that's a very profitable market, so it's not much of a downside to those who are just looking to make money. Sure, it would be a soulless, devoid experience, but today's pop music is pretty much like that anyway.
Other music couldn't be handled that way though. Many styles outside of pop music still need to actually have some soul to them for people to accept it, whether it's metal, classical, jazz, etc. Plus, part of what makes it enjoyable is the human element of it. If you have someone playing an amazing guitar solo, part of the enjoyment comes from knowing a person worked hard developing the skill to play it, and was able to put the emotion into it. A machine playing the same solo isn't impressive. The same goes for any other instrument, and for vocals. A human hitting A5 is incredible; a computer doing so isn't.
There's also the emotion of the artist performing that you get with genuine music, that you couldn't get with a machine. Whether it's seeing the artist genuinely enjoying themselves while playing, or seeing them truly express the emotion of the song, you need an actual person doing it.
Crowd interaction is also pretty important outside of pop music. You need people for good crowd interaction.
tl;dr
AI could work for pop music where people just want to go and watch essentially a pre-rendered performance, but won't work for other genres.
"Fourteen year old Chia Pet McKenzie is chosen by the group to go to Tokyo" huh? I suppose if I read the book it would make sense.
Yep.
And I still want a Sandbenders laptop.
I think you're really getting on a musical high horse here. I think, if an AI is capable of generating pop music, it's capable of generating "classical" music, jazz music, and a lot of others.
Now, you're absolutely right that it wouldn't be that impressive in seeing a performance of some of those genres of music, because of the perfect nature of mechanical ability. In the case of AI-created music, the amazing thing is the composition itself.
This. It's much easier and cheaper for a producer to simply hire an actual human being to be the "pop star" than to invest in creating some CGI pop star.
There are already more than enough digital studio tricks to turn a mediocre singer into a pop singer (and vocals that sound blatantly synthetic are become more accepted as well), lip-synching has become more and more acceptable during live performances, and nobody really expects modern pop stars to write or compose their own songs.
Someone needs to give Milli Vanilli their Grammy back.
It's not quite the same thing, but there are "virtual" musical artists like the Gorillaz or Lumi of Genki Rockets, who aren't real people but are portrayed musically by actual people, while all their appearances in the multimedia are manufactured.
Rock Band DLC | GW:OttW - arrcd | WLD - Thortar
Weird!
For what it's worth, I think arguing that 'degree of difficulty associated with performance' is a pretty empty way of juding music. Lots of stuff is hard as hell to perform, but that's quite different from it being good.