Inspired by the "Lesser of two Evils" thread, I have the following question:
Do you think it is possible - given enough time and computing and audio horsepower - to create a computer program that can play tunes from the dots with the same beauty as a good player? In other words, if it was then recorded onto a CD, most of us wouldn't be able to tell that it wasn't a skilled traditional musician? A super-steroidal ABCMUS program, if you will, with codified instructions as to how and where you might ornament (doing it differently each time through, of course) how to create variations, and so on.
Yes, of course. Just as it's possible to create a machine to run faster than a cheetah, be stronger than a gorilla or look uglier than Karl Malden. Not sure who it would appeal to, though.
I'm quite sure that an expert system could be created. It would need rules about phrasing, speed, ornamentation etc., but I can't see any reason why it couldn't be done. It would certainly OPERATE differently from a live player, just like computer chess programs work differently from grandmasters. In that context, I'd like to bet that nobody here could do better than draw the occasional game with one of those beasts.
It's a very good question, because it's one that goes to the heart of the nature of human thinking. In the end, all creativity is algorithmic- it builds on what it has done before, and on what others did before that. But it's not mechanical, any more than genetic evolution is mechanical. The interaction of the mutation (the idea if you like), the internal environment, and the ecology that surrounds it (the session in this case) means that the outcome is totally unpredictable.
As an aside, I think that it's the intimate, direct testing of ideas against other equals with similar ideas that makes the session such an exciting dynamic process.
Whether the machine feels anything is another matter- it's often hard enough to tell if another human being has feeling.
Star treck went there, with having the Data character play mozart. But the answere in the real world is "not a cat in hell's chance". For starters, You can't even get a sythersizer to sound like an acoustic instrument, never mind a sequencer to play like a human.
Gord says, "Yes, of course. Just as it's possible to create a machine to run faster than a cheetah". Show me a machine that can run as fast as a cheetah over uneven scrub? Sure you can make a machine with wheels and rubber tyres and suspension that can go well over 70mph but it's not running is it. And even this wheeled thing can't turn a gazelle like a cheetah. Show me a machine that san catch a gazelle? Come off it
"all creativity is algorithmic" I would disagree with this statement Paul the Church-Turing Thesis is not proven and work done on non-linear systems suggest that other forms of non-symbolic computation may be involved in the CNS. http://www.ece.utexas.edu/~werner/Neural_computation.html
So could a Machine play the music well?
I think so but it would probably have to be a hardware system implementing coupled non-linear (relaxation) oscillators showing the various routes to Chaos.
And then there's the question of consciousness but I can see people’s eyes glazing over so I'll leave that for a later date.
I've read recently an article about a (German?) researcher which studied the style of different piano players from their recordings. I guess using some methods of neural networks pattern recoginition and comparing the dots with what was played actually. One application they mentioned is to analyze old recordings and then" rerelease" them with a new recorded piano played by the virtual piano player .....
I’m not sure about this being the ULTIMATE evil. I mean, what if someone were to create a completely artificial singer. Airbrush out the acne, put the voice in tune, even give it a cheeky grin. Faceless automatons could compose the music. It could be called Robbie Williams.
PP, we're probably talking about algorithms operating on different scales. What I meant was the simple mutate; test; accept/reject process.
The availability of modifications, and the nature of the tests to which they are subjected, is (possibly) a different kettle of fish(*), and is what I meant by the internal environment . In other words, the loop may be iterated thousands of times before anyone even becomes aware of it, and in that many generations a lot of evolution can occur.
(*) In many years, I've never cooked fish in a kettle, partly because of the problem of controlling the oil temperature.
I'm with Gords... it is possible. Or if it isn't, it will be soon. It's essentially the same funciton as a speech synthesiser isn't it? The software/AI/algorithm parses the inputted ABC, processes the rhythm so that slight irregularity profiles are incorporated, selects pre-recorded or processed samples from its sound bank, re-processes to smooth over anomolous glitches and there you have it. It would just need a LOT of time and effort to set up such a programme. And a lot of funding too. But believe me, they'll get around to it sooner rather than later.
And I wouldn't mind. I mean, seriously, who'd use it? I would, so that I could play duets with my laptop when I was bored. But who else, and for what reason? Record companies? Please... Publicans? Maybe... but they may as well stick on a CD.
Yes, it is essentially the same function as a speech synthersiser. But with all the billions that have gone into that, they are stills oceans away from creating one that may fool you into thinking it was a real recorded voice. And light years away from creating one that would fool you in conversation.
And I'd suggest that playing music is even more cognitivley complex that speech
fair enough. But you're thinking too laterally. Literally! Oceans away? Light years? Those are descriptions of distance, whereas I'm arguing that it's only a matter of time... I'll bet you an Altyrian megadollar that we'll have flawless speech synthesisers within 20 years, probably way sooner, and music synthesisers soon after.
My confidence in this is from remembering where speech synthesis was 20 years ago, and knowing where it is today (have you messed around with the different speech synths on OSX? the improvement), noting the extent and rate of improvement. It's one of them whatsit curves... starts off slow and low, and then picks up speed and angle until it approaches but never "reaches" infinity. Or perfect human playing, in this case.
C'mon, an altyrian megadollar's on the table. Who'll take my bet?
Yes, an exponential curve, e to the power of whatever. And I haven't tried the speech on OSX yet, must try that. It was always fun to get the OS9 one to read out a long list of the most expresivly sexual swearwords you could think of.
But a good example was the Roland human rythm composer, remember that one? It's subtle quantising and randomness really was quite effective. But remember that that played samples, it didn't create the sounds. And there's the obvious comparison to how easy it is for a machine to emulate a drummer, but when it comes to a musician?
The bet's on. One altyrian megadollar it is. (index linked?)
Speaking of trademarkes, when I was at the new Irish Pub at Disneyworld, the souveneirs with the pub's name on them were marked "RAGLAN ROAD (TM)." I guess you folks back in Ireland will have to rename your road....the Mouse owns the name now!
(I will now duck and hide under my desk.)
Perhaps it is possible to make a machine that would play a piece of music correctly, maybe even beautifully--but wouldn't that be missing the point? The real beauty of a masterful musical performance is that it's an achievement of a human being, one that other humans can enjoy and be inspired by.
I recently saw a rock band that obviously used a sampler for the keyboard and bass parts of all the songs--the keyboardist just pantomimed, with a smirk on his face. Sure, there were no mistakes, but it had no soul. The whole thing, for me, was like the face of a beautiful woman--but one with dead blank spaces where the eyes (windows of the soul, right?) ought to be.
Yes, mickray. That's why it's the ultimate evil. But the thought is intriguing in its implications.
Suppose a masterfully executed program like this generated a set of fiddle tunes in the style of Michael Coleman, and some unscrupulous person issued this as a "lost" Coleman recording (the bad piano would be easy to add, I think.) And suppose he fooled all of the people for some of the time, and they derived great enjoyment altogether from listening to the great Michael one more time. And many session players would maybe learn the set "by ear" from the CD because that was really so much more authentic than reading the dots.
"The real beauty of a masterful musical performance is that it's an achievement of a human being, one that other humans can enjoy and be inspired by."
Really? I think you'd get some argument on that. For me, at least, beautiful music is its own reward.
"You can't even get a synthesizer to sound like an acoustic instrument..."
Have you listened to an electronic woodwind? You probably have, unwittingly.
As someone with a PhD in AI, I think its quite possible to do this -- even trivial in some senses. Playing beautiful music is very difficult for people, but much of what makes it so difficult for people is easy for a computer. All the mechanics of playing, the memorizing of tunes, keeping a steady tempo, etc., is mere rote and easy to program. Building a computer that could play at a journeyman level would not be terribly difficult. I suspect that ABCmus with better sound output and a few stylistic phrasings would fool most naive listeners. Of course, the last part from journeyman on would be far more difficult.
The more interesting challenges to me lie in ensemble playing and composing. My doctoral work was on a program to write stories about King Arthur -- and in a similar way I think the problem of composing music is much more difficult and interesting than playing music. Ensemble playing has some challenging aspects as well -- the feedback loop and so on.
I'm with you on the "beautiful music is its own reward" bit. Except if I'm hankering to play, that is. Then the robot better move his capacitors outta my seat.
Music is more than physics and mathematics, or at least it should be. Sure, you can analyze the details of a particular player's performance of a piece to the nth degree, but all you'll get is a very detailed description of how he or she played it that time, or the few times it was recorded. No computer can guess how he or she will interpret the piece the next time, because no machine can have the emotional and life-experience complexities of an adult human being. Not any machines we have developed to date, anyway. Maybe we never will--human beings are really really complicated. I doubt that you can break down the human psyche into a set of "yes/no" instructions, no matter how many lines of code you write.
I think its a bit like playing Quake. Sure, the bots are getting better and better. They can be harder to beat that your mate. But nothing about bots is as good as fragging your best mate. All the banter and abuse is so much more fun than the bot non-messages. Likewise, playing in a session with real live people you have built a relationship with will always beat the atmosphere of a session bot. The alternative is sophisticated onanism.
I think you'll find there is such a soulless machine in operation, which can operate from 'codified instructions'. It's called Ceoltas Comhaltairi Eireann
Oooooooooooh...... Ye'll hae the 'Comhaltas Police' after ye now copo, for sure! You can run, but you cannot hide! Just listen out for their sniffer dogs. Time to get out that false beard!
Of course it is possible. At university we wrote a program that simulated human writing, and it worked extremely well. You could set all sorts of parameters, like the weight of the pen, the strength of the hand and the mood that the writer is in. You could set it so that the mood changes as the writing continues and the hand could become tired when a lot is written or if writing is fast. My room mate and I handed in many papers written in this way, and nobody noticed. His was set to "creative, inteligent and angry" while mine was set to "anal retentive and slightly effeminate". (Not because I was, I just liked the handwriting those settings produced.) That was in 1988.
If there was a need for such a program, I would be rich. But there isn't, and that's the key.
Music is driven by the musician not the listener. People are willing to put in a lot of work when they feel pasionate about something. Nobody is as pasionate about listening to music as the musician is about creating or performing it.
So yes, it can already be done, but nobody is willing to put in the work that it will require.
And another thing: Music generated by a computer will only lack "soul" if the program is written incorrectly. I've heard the London Philharmonic Orchestra live. They played flawlessly. I was extremely bored and very dissapointed. What you call "soul" is simply the collected affects of glitches, errors, human inconsistancy and distraction. When someone plays something that really moves you, you cannot tell whether they're thinking about misty meadows or worrying about this year's tax returns. You just assume.
They've just created a rat brain in a jar that can pilot a fighter jet in a simulator. Word is, a new generation of computers with little rat brains inside is on the horizon. I think a rat-brain computer compatible program would be better at playing tunes than, say, MIDI is.
But then, who would you rather see in concert, Alasdair Fraser or a brain in a jar?
(Actually, come to think of it, I'd probably go for the brain in the jar myself...)
People play music because they enjoy it, not to achieve some predetermined end.
Having a computer do it for you takes away the enjoyment, no matter how good or "soulful" the music resulting is.
This post was written by a human. Not by a computer programmed to write like a human. No really. Not like that so-called "Dow" up there. He's not real at all. Pretty obvious.
Might as well add my tuppence worth - to me music is not something which is exact. Did you know that many years ago, Tom Morrow of Dervish did an investigation into the styles of many top fiddle players an found that ALL of them played with certain idiosyncasies - such as a slightly sharpened note here and there, or rhythmic variations.
Donal Lunny doesn't play strictly on or off beat, but it sounds great!
Computers and synths have been around for a long time. In the 1970s electronic music was popular for a while until people realised that the essential human factor was missing. With further advances in technology there is now the notion of programming in "variations" to make electronic music interesting. With all due respect to our resident PhD in AI, the commercial breakthrough has yet to be made? Whether or not it is feasible or not, that';s another question. "Expert systems" have been around for a long time but they have yet to take over.....
Even for a simple thing like drums, drum machines still don't give the same "feel" as a live drummer. Ace perfectionists Steely Dan have been known to spend weeks with various drummers trying to get the sound they want - don't you think that if there was a better drum machine they would use it (and I happen to know they have tried to create one).
My conclusion - people want their music created by people and I don't see it changing any time soon. However, technology can certainly provide major crutches both in a studio and a live environment to enhance and improve nature.
The Ultimate Evil
The Ultimate Evil
Inspired by the "Lesser of two Evils" thread, I have the following question:
Do you think it is possible - given enough time and computing and audio horsepower - to create a computer program that can play tunes from the dots with the same beauty as a good player? In other words, if it was then recorded onto a CD, most of us wouldn't be able to tell that it wasn't a skilled traditional musician? A super-steroidal ABCMUS program, if you will, with codified instructions as to how and where you might ornament (doing it differently each time through, of course) how to create variations, and so on.
# Posted on December 6th 2005 by grego
Re: The Ultimate Evil
I'm sorry, but I just can't get behind that
# Posted on December 6th 2005 by Just a person
Re: The Ultimate Evil
Unfortunately, it's probably only a matter of time. Please don't throw down the challenge because someone will feel they have to rise to it!!
# Posted on December 6th 2005 by Tarrantella
Re: The Ultimate Evil
Yes, of course. Just as it's possible to create a machine to run faster than a cheetah, be stronger than a gorilla or look uglier than Karl Malden. Not sure who it would appeal to, though.
# Posted on December 6th 2005 by Gords
Re: The Ultimate Evil
I'm quite sure that an expert system could be created. It would need rules about phrasing, speed, ornamentation etc., but I can't see any reason why it couldn't be done. It would certainly OPERATE differently from a live player, just like computer chess programs work differently from grandmasters. In that context, I'd like to bet that nobody here could do better than draw the occasional game with one of those beasts.
It's a very good question, because it's one that goes to the heart of the nature of human thinking. In the end, all creativity is algorithmic- it builds on what it has done before, and on what others did before that. But it's not mechanical, any more than genetic evolution is mechanical. The interaction of the mutation (the idea if you like), the internal environment, and the ecology that surrounds it (the session in this case) means that the outcome is totally unpredictable.
As an aside, I think that it's the intimate, direct testing of ideas against other equals with similar ideas that makes the session such an exciting dynamic process.
Whether the machine feels anything is another matter- it's often hard enough to tell if another human being has feeling.
# Posted on December 6th 2005 by LastToFinish
Re: The Ultimate Evil
Grego, cut it out! You’re scaring me!
# Posted on December 6th 2005 by Bob himself
Re: The Ultimate Evil
Star treck went there, with having the Data character play mozart. But the answere in the real world is "not a cat in hell's chance". For starters, You can't even get a sythersizer to sound like an acoustic instrument, never mind a sequencer to play like a human.
Gord says, "Yes, of course. Just as it's possible to create a machine to run faster than a cheetah". Show me a machine that can run as fast as a cheetah over uneven scrub? Sure you can make a machine with wheels and rubber tyres and suspension that can go well over 70mph but it's not running is it. And even this wheeled thing can't turn a gazelle like a cheetah. Show me a machine that san catch a gazelle? Come off it
# Posted on December 6th 2005 by ...
Re: The Ultimate Evil
No. Nay. Nyet. Nein.
# Posted on December 6th 2005 by cathrynb
Re: The Ultimate Evil
"all creativity is algorithmic" I would disagree with this statement Paul the Church-Turing Thesis is not proven and work done on non-linear systems suggest that other forms of non-symbolic computation may be involved in the CNS.
http://www.ece.utexas.edu/~werner/Neural_computation.html
So could a Machine play the music well?
I think so but it would probably have to be a hardware system implementing coupled non-linear (relaxation) oscillators showing the various routes to Chaos.
And then there's the question of consciousness but I can see people’s eyes glazing over so I'll leave that for a later date.
PP
# Posted on December 6th 2005 by Pied Piper
Re: The Ultimate Evil
I've read recently an article about a (German?) researcher which studied the style of different piano players from their recordings. I guess using some methods of neural networks pattern recoginition and comparing the dots with what was played actually. One application they mentioned is to analyze old recordings and then" rerelease" them with a new recorded piano played by the virtual piano player .....
# Posted on December 6th 2005 by swisspiper
Re: The Ultimate Evil
I’m not sure about this being the ULTIMATE evil. I mean, what if someone were to create a completely artificial singer. Airbrush out the acne, put the voice in tune, even give it a cheeky grin. Faceless automatons could compose the music. It could be called Robbie Williams.
# Posted on December 6th 2005 by Gords
Re: The Ultimate Evil
PP, we're probably talking about algorithms operating on different scales. What I meant was the simple mutate; test; accept/reject process.
The availability of modifications, and the nature of the tests to which they are subjected, is (possibly) a different kettle of fish(*), and is what I meant by the internal environment . In other words, the loop may be iterated thousands of times before anyone even becomes aware of it, and in that many generations a lot of evolution can occur.
(*) In many years, I've never cooked fish in a kettle, partly because of the problem of controlling the oil temperature.
# Posted on December 6th 2005 by LastToFinish
Re: The Ultimate Evil
You mean, if you play The Loop thousands of times it changes?
# Posted on December 6th 2005 by Trevor Jennings
Re: The Ultimate Evil
I'm with Gords... it is possible. Or if it isn't, it will be soon. It's essentially the same funciton as a speech synthesiser isn't it? The software/AI/algorithm parses the inputted ABC, processes the rhythm so that slight irregularity profiles are incorporated, selects pre-recorded or processed samples from its sound bank, re-processes to smooth over anomolous glitches and there you have it. It would just need a LOT of time and effort to set up such a programme. And a lot of funding too. But believe me, they'll get around to it sooner rather than later.
And I wouldn't mind. I mean, seriously, who'd use it? I would, so that I could play duets with my laptop when I was bored. But who else, and for what reason? Record companies? Please... Publicans? Maybe... but they may as well stick on a CD.
# Posted on December 6th 2005 by Q
Re: The Ultimate Evil
Yes, it is essentially the same function as a speech synthersiser. But with all the billions that have gone into that, they are stills oceans away from creating one that may fool you into thinking it was a real recorded voice. And light years away from creating one that would fool you in conversation.
And I'd suggest that playing music is even more cognitivley complex that speech
# Posted on December 6th 2005 by ...
Re: The Ultimate Evil
fair enough. But you're thinking too laterally. Literally! Oceans away? Light years? Those are descriptions of distance, whereas I'm arguing that it's only a matter of time... I'll bet you an Altyrian megadollar that we'll have flawless speech synthesisers within 20 years, probably way sooner, and music synthesisers soon after.
My confidence in this is from remembering where speech synthesis was 20 years ago, and knowing where it is today (have you messed around with the different speech synths on OSX? the improvement), noting the extent and rate of improvement. It's one of them whatsit curves... starts off slow and low, and then picks up speed and angle until it approaches but never "reaches" infinity. Or perfect human playing, in this case.
C'mon, an altyrian megadollar's on the table. Who'll take my bet?
# Posted on December 6th 2005 by Q
Re: The Ultimate Evil
er, please excuse occasional confusion in my post above. A black hole swallowed my syntax.
# Posted on December 6th 2005 by Q
Re: The Ultimate Evil
Call me a Luddite but I certainly hope not.
# Posted on December 6th 2005 by wormdiet
Re: The Ultimate Evil
Yes, an exponential curve, e to the power of whatever. And I haven't tried the speech on OSX yet, must try that. It was always fun to get the OS9 one to read out a long list of the most expresivly sexual swearwords you could think of.
But a good example was the Roland human rythm composer, remember that one? It's subtle quantising and randomness really was quite effective. But remember that that played samples, it didn't create the sounds. And there's the obvious comparison to how easy it is for a machine to emulate a drummer, but when it comes to a musician?
The bet's on. One altyrian megadollar it is. (index linked?)
# Posted on December 6th 2005 by ...
Re: The Ultimate Evil
Quantising, that's the word I was looking for! I had to delete a whole section of my previous post cos I couldn't remember what the damn word was
excellent. I'll see you back here in 20 then. Index, glossary, bibliography... whatever link you want - just as it's not missing!
and after you've handed over the money, we can have a session in the kitchen, with the "Frankie Gavin™ fridge and the Matt Molloy™ microwave!
# Posted on December 6th 2005 by Q
Re: The Ultimate Evil
Speaking of trademarkes, when I was at the new Irish Pub at Disneyworld, the souveneirs with the pub's name on them were marked "RAGLAN ROAD (TM)." I guess you folks back in Ireland will have to rename your road....the Mouse owns the name now!
(I will now duck and hide under my desk.)
# Posted on December 6th 2005 by AlBrown
Re: The Ultimate Evil
Al, that isn't a joke, is it?
# Posted on December 6th 2005 by Trevor Jennings
Re: The Ultimate Evil
Perhaps it is possible to make a machine that would play a piece of music correctly, maybe even beautifully--but wouldn't that be missing the point? The real beauty of a masterful musical performance is that it's an achievement of a human being, one that other humans can enjoy and be inspired by.
I recently saw a rock band that obviously used a sampler for the keyboard and bass parts of all the songs--the keyboardist just pantomimed, with a smirk on his face. Sure, there were no mistakes, but it had no soul. The whole thing, for me, was like the face of a beautiful woman--but one with dead blank spaces where the eyes (windows of the soul, right?) ought to be.
# Posted on December 6th 2005 by John Galt
Re: The Ultimate Evil
Yes, mickray. That's why it's the ultimate evil. But the thought is intriguing in its implications.
Suppose a masterfully executed program like this generated a set of fiddle tunes in the style of Michael Coleman, and some unscrupulous person issued this as a "lost" Coleman recording (the bad piano would be easy to add, I think.) And suppose he fooled all of the people for some of the time, and they derived great enjoyment altogether from listening to the great Michael one more time. And many session players would maybe learn the set "by ear" from the CD because that was really so much more authentic than reading the dots.
How evil would that be?
# Posted on December 6th 2005 by grego
Re: The Ultimate Evil
"The real beauty of a masterful musical performance is that it's an achievement of a human being, one that other humans can enjoy and be inspired by."
Really? I think you'd get some argument on that. For me, at least, beautiful music is its own reward.
"You can't even get a synthesizer to sound like an acoustic instrument..."
Have you listened to an electronic woodwind? You probably have, unwittingly.
As someone with a PhD in AI, I think its quite possible to do this -- even trivial in some senses. Playing beautiful music is very difficult for people, but much of what makes it so difficult for people is easy for a computer. All the mechanics of playing, the memorizing of tunes, keeping a steady tempo, etc., is mere rote and easy to program. Building a computer that could play at a journeyman level would not be terribly difficult. I suspect that ABCmus with better sound output and a few stylistic phrasings would fool most naive listeners. Of course, the last part from journeyman on would be far more difficult.
The more interesting challenges to me lie in ensemble playing and composing. My doctoral work was on a program to write stories about King Arthur -- and in a similar way I think the problem of composing music is much more difficult and interesting than playing music. Ensemble playing has some challenging aspects as well -- the feedback loop and so on.
-- Scott
# Posted on December 6th 2005 by srt19170
Re: The Ultimate Evil
Woo Woo! Go Scott!
I'm with you on the "beautiful music is its own reward" bit. Except if I'm hankering to play, that is. Then the robot better move his capacitors outta my seat.
"Trivial" - I love it!
# Posted on December 6th 2005 by Q
Re: The Ultimate Evil
And do wath? A traditional rave party?
# Posted on December 7th 2005 by pitnekit
Re: The Ultimate Evil
Music is more than physics and mathematics, or at least it should be. Sure, you can analyze the details of a particular player's performance of a piece to the nth degree, but all you'll get is a very detailed description of how he or she played it that time, or the few times it was recorded. No computer can guess how he or she will interpret the piece the next time, because no machine can have the emotional and life-experience complexities of an adult human being. Not any machines we have developed to date, anyway. Maybe we never will--human beings are really really complicated. I doubt that you can break down the human psyche into a set of "yes/no" instructions, no matter how many lines of code you write.
# Posted on December 7th 2005 by John Galt
Re: The Ultimate Evil
I think its a bit like playing Quake. Sure, the bots are getting better and better. They can be harder to beat that your mate. But nothing about bots is as good as fragging your best mate. All the banter and abuse is so much more fun than the bot non-messages. Likewise, playing in a session with real live people you have built a relationship with will always beat the atmosphere of a session bot. The alternative is sophisticated onanism.
# Posted on December 7th 2005 by NeilBarr
Re: The Ultimate Evil
I think you'll find there is such a soulless machine in operation, which can operate from 'codified instructions'. It's called Ceoltas Comhaltairi Eireann
# Posted on December 7th 2005 by copo24
Re: The Ultimate Evil
Oooooooooooh...... Ye'll hae the 'Comhaltas Police' after ye now copo, for sure! You can run, but you cannot hide! Just listen out for their sniffer dogs. Time to get out that false beard!
# Posted on December 7th 2005 by Ptarmigan
Re: The Ultimate Evil
Of course it is possible. At university we wrote a program that simulated human writing, and it worked extremely well. You could set all sorts of parameters, like the weight of the pen, the strength of the hand and the mood that the writer is in. You could set it so that the mood changes as the writing continues and the hand could become tired when a lot is written or if writing is fast. My room mate and I handed in many papers written in this way, and nobody noticed. His was set to "creative, inteligent and angry" while mine was set to "anal retentive and slightly effeminate". (Not because I was, I just liked the handwriting those settings produced.) That was in 1988.
If there was a need for such a program, I would be rich. But there isn't, and that's the key.
Music is driven by the musician not the listener. People are willing to put in a lot of work when they feel pasionate about something. Nobody is as pasionate about listening to music as the musician is about creating or performing it.
So yes, it can already be done, but nobody is willing to put in the work that it will require.
And another thing: Music generated by a computer will only lack "soul" if the program is written incorrectly. I've heard the London Philharmonic Orchestra live. They played flawlessly. I was extremely bored and very dissapointed. What you call "soul" is simply the collected affects of glitches, errors, human inconsistancy and distraction. When someone plays something that really moves you, you cannot tell whether they're thinking about misty meadows or worrying about this year's tax returns. You just assume.
# Posted on December 7th 2005 by Shrog
Re: The Ultimate Evil
They've just created a rat brain in a jar that can pilot a fighter jet in a simulator. Word is, a new generation of computers with little rat brains inside is on the horizon. I think a rat-brain computer compatible program would be better at playing tunes than, say, MIDI is.
But then, who would you rather see in concert, Alasdair Fraser or a brain in a jar?
(Actually, come to think of it, I'd probably go for the brain in the jar myself...)
# Posted on December 7th 2005 by Kerri Brown
Re: The Ultimate Evil
Never mind a computer - my mobile phone could make a better job of playing a tune than some of the Irish musicians I've played with
# Posted on December 7th 2005 by Dr. Dow
Re: The Ultimate Evil
People play music because they enjoy it, not to achieve some predetermined end.
Having a computer do it for you takes away the enjoyment, no matter how good or "soulful" the music resulting is.
This post was written by a human. Not by a computer programmed to write like a human. No really. Not like that so-called "Dow" up there. He's not real at all. Pretty obvious.
# Posted on December 7th 2005 by Bren
Re: The Ultimate Evil
Might as well add my tuppence worth - to me music is not something which is exact. Did you know that many years ago, Tom Morrow of Dervish did an investigation into the styles of many top fiddle players an found that ALL of them played with certain idiosyncasies - such as a slightly sharpened note here and there, or rhythmic variations.
Donal Lunny doesn't play strictly on or off beat, but it sounds great!
Computers and synths have been around for a long time. In the 1970s electronic music was popular for a while until people realised that the essential human factor was missing. With further advances in technology there is now the notion of programming in "variations" to make electronic music interesting. With all due respect to our resident PhD in AI, the commercial breakthrough has yet to be made? Whether or not it is feasible or not, that';s another question. "Expert systems" have been around for a long time but they have yet to take over.....
Even for a simple thing like drums, drum machines still don't give the same "feel" as a live drummer. Ace perfectionists Steely Dan have been known to spend weeks with various drummers trying to get the sound they want - don't you think that if there was a better drum machine they would use it (and I happen to know they have tried to create one).
My conclusion - people want their music created by people and I don't see it changing any time soon. However, technology can certainly provide major crutches both in a studio and a live environment to enhance and improve nature.
# Posted on December 7th 2005 by lysaghtm