Emily Howell
http://www.slate.com/id/2254232/
computer (assisted) generated music
-
- KVRAF
- 2217 posts since 15 Jul, 2003
-
- KVRian
- 1084 posts since 12 Sep, 2008 from Your basement
I had the pleasure of listening to a lecture delivered by Professor Cope. Here's my thoughts on computer-assisted composing...wrench45us wrote:Emily Howell
http://www.slate.com/id/2254232/
It is a springboard process by which you fill your "musical reservoir" with ideas and possibilities. This is not unlike listening to a bunch of recordings or performances and becoming inspired to write an original composition.
I think we've all had the experience of listening to half a song in a coffee shop and were given an idea on a brand new piece of music or a song.
Emily is not a substitute for the compositional process...it's a way of sparking ideas and stimulating possibility. I think it's great. Human composers will always be valued for their ability to creatively connect phrases. Where the spark of ideas come from is not the thing of value, nor has it ever been.
Anyway thanks for posting that, it's interesting to think about isn't it?
-
- KVRist
- 350 posts since 11 May, 2008
My Master thesis was about this issues (case studies: Cope and Essl Lexikon sonate). And now my Phd Thesis goes around creating a system to generate a particular kind of music.
I'm very very interested in this branch of research, and basically what I have to say is the following:
Composition is a way to solve a problem. When you compose, you are either filling gaps, or adding another chain to the link.
You have one of the two: you have a general idea of a whole, and then you have to know "what I need to put there to reach this structure, or to obtain this function or effect", or "I want to go from here. What's next that will make sense to what lies behind?".
Well, humans, and specially creative humans are excellent in finding ways to do this. Other times they got blank.
However in every period, theory always found "rules" and tricks, and algorithms to make it easier.
On tonal music systems, we already know that after a ii7 a V7 will work just fine. And that after a V7 a I will do marvelous. It has become a rule.
On every systems we can find patterns and rules.
And computers are very good at following rules, or even finding them by giving them a bunch of songs.
If you feed a computer 500 songs and ask him to compare them and find patters it can be very easy for it to do it.
Computers are wonderfull at making repetitive tasks quick and harmless. While humans might take ages.
So, If we can have a computer making this tasks for us, why not? And they might even surprise us!...
Now...
The real revolution would be when a computer is able to LISTEN to what it does, and then reason about it and modify it accordingly. That's the big step that's yet to be done.
I'm very very interested in this branch of research, and basically what I have to say is the following:
Composition is a way to solve a problem. When you compose, you are either filling gaps, or adding another chain to the link.
You have one of the two: you have a general idea of a whole, and then you have to know "what I need to put there to reach this structure, or to obtain this function or effect", or "I want to go from here. What's next that will make sense to what lies behind?".
Well, humans, and specially creative humans are excellent in finding ways to do this. Other times they got blank.
However in every period, theory always found "rules" and tricks, and algorithms to make it easier.
On tonal music systems, we already know that after a ii7 a V7 will work just fine. And that after a V7 a I will do marvelous. It has become a rule.
On every systems we can find patterns and rules.
And computers are very good at following rules, or even finding them by giving them a bunch of songs.
If you feed a computer 500 songs and ask him to compare them and find patters it can be very easy for it to do it.
Computers are wonderfull at making repetitive tasks quick and harmless. While humans might take ages.
So, If we can have a computer making this tasks for us, why not? And they might even surprise us!...
Now...
The real revolution would be when a computer is able to LISTEN to what it does, and then reason about it and modify it accordingly. That's the big step that's yet to be done.
Play fair and square!
-
- KVRian
- 1084 posts since 12 Sep, 2008 from Your basement
What, in your opinion, is required for a computer to do that? A high level of artificial intelligence? How many years do you think we are away from that level of technology?Musicologo wrote: The real revolution would be when a computer is able to LISTEN to what it does, and then reason about it and modify it accordingly. That's the big step that's yet to be done.
-
- KVRist
- 350 posts since 11 May, 2008
I don't know. Music is a human phenomena. Computers only produce sound not music. We are the ones who make music.
For a computer to make music (listen to the sounds they produce and make sense of that), they would have to be able to make a similar process to what we do when we hear sounds as music. It's a neurological thing I guess.
But as far as I know, noone ever discovered what makes that happen, so we are not able to mimic that in a machine.
For a computer to make music (listen to the sounds they produce and make sense of that), they would have to be able to make a similar process to what we do when we hear sounds as music. It's a neurological thing I guess.
But as far as I know, noone ever discovered what makes that happen, so we are not able to mimic that in a machine.
Play fair and square!
-
- Banned
- 12367 posts since 30 Apr, 2002 from i might peeramid
http://en.wikipedia.org/wiki/Metaprogramming
afaik timothy leary coined the term as a fancy way of communicating the idea of actually being responsible for and having control over one's own perceptions and reactions. interesting how modern culture seems to have coopted it for purely mechanistic, extricative processes designed to accomplish an end within the scope of enterprise.
isn't music first what our culture tells us it is, then next what we decide it is, and perhaps, finally, what younger people aren't able to appreciate? given that, i ask what's left? what isn't music? perhaps the most accurate definition i've seen is 'music is time'.
can computers generate and regenerate events in time? perhaps if we're big enough not to tell them when they aren't doing it properly..
afaik timothy leary coined the term as a fancy way of communicating the idea of actually being responsible for and having control over one's own perceptions and reactions. interesting how modern culture seems to have coopted it for purely mechanistic, extricative processes designed to accomplish an end within the scope of enterprise.
isn't music first what our culture tells us it is, then next what we decide it is, and perhaps, finally, what younger people aren't able to appreciate? given that, i ask what's left? what isn't music? perhaps the most accurate definition i've seen is 'music is time'.
can computers generate and regenerate events in time? perhaps if we're big enough not to tell them when they aren't doing it properly..
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.
-
- Banned
- 6129 posts since 9 Oct, 2007 from an inharmonious society
I assume space to be time, and music a sound which is a wave.xoxos wrote: 'music is time'.
but I see it as more poetic and creative the way you stated it.
I do enjoy using some computer generated software sometimes to get ideas.
I did like Lexicon Sonate, and the developers other generative software programs, which are genius imo.
I do use Thesys quite often for great ideas, that can be more controlled by the user, but can still do a great degree of computer generated actions as well.
-
- KVRAF
- Topic Starter
- 2217 posts since 15 Jul, 2003
the manner of computer programming would have a great deal to do with the 'sucress' of the output. I'm sure Cope has learned a great deal including perhaps different models and ways to think about and 'quantify' music.
It would be a mistake to think of computer generated music as having sprung forth from a computer by itself without a great deal of human programming -- every compositional decison would be based on logic rules the programmer has established.
so one developer's computer music is an individual creative process as involved and intricate as direct music composition.
The fact that Cope derived rules from various composition styles is what allows the listener to supposedly easily identify the influences in the output. I see that as a measure of his skill in programming -- that and the ability to tweak and evolve the program to produce more pleasing compositions.
I find it all rather fascinating because music is based on certain rules and logic choices that have evolved with the ages and splintered into a confusing number of genre and sub-genre.
Having said all that I still think any given music is more or less a structure for the performer(s) to arrange and animate with emotion and that's what people relate to.
for example, I think it would be relatively easy to generate a computer program to create a walking bass pattern from a given chord progression. The rules are well known and understood, but a player adds all sorts of shifting nuances that would be a lot meore difficult for a computer to model and emulate.
It would be a mistake to think of computer generated music as having sprung forth from a computer by itself without a great deal of human programming -- every compositional decison would be based on logic rules the programmer has established.
so one developer's computer music is an individual creative process as involved and intricate as direct music composition.
The fact that Cope derived rules from various composition styles is what allows the listener to supposedly easily identify the influences in the output. I see that as a measure of his skill in programming -- that and the ability to tweak and evolve the program to produce more pleasing compositions.
I find it all rather fascinating because music is based on certain rules and logic choices that have evolved with the ages and splintered into a confusing number of genre and sub-genre.
Having said all that I still think any given music is more or less a structure for the performer(s) to arrange and animate with emotion and that's what people relate to.
for example, I think it would be relatively easy to generate a computer program to create a walking bass pattern from a given chord progression. The rules are well known and understood, but a player adds all sorts of shifting nuances that would be a lot meore difficult for a computer to model and emulate.
-
- Banned
- 6129 posts since 9 Oct, 2007 from an inharmonious society
I have a lot of links to this kind of software.
Many seem to be mac only, but I'll sort them and post em here later.
I also have Chill, which is mac only.
I have used it to get the midi which I then edited and built a tune mixing parts, and adding my own playing to come up with something. The outcome of which really didn't work so great, but it was a learning experience just the same.
Many seem to be mac only, but I'll sort them and post em here later.
I also have Chill, which is mac only.
I have used it to get the midi which I then edited and built a tune mixing parts, and adding my own playing to come up with something. The outcome of which really didn't work so great, but it was a learning experience just the same.
