Music Composition Decision Path

[mer]

I liked reading the list - amazing to see how many decisions are being made, often "behind the scenes" that result in a piece of music. Lately the question I'm concious of asking and answering first about a piece is whether it will be in 3 or 2. Coming from a fiddle tune background, I find myself deciding first if it is a waltz (3), a reel (2 or 4), or a jig (2 main beats each with 3 smaller beats each), or something else. When I started seriously composing in 2000 I wrote all in 3, either waltzes or jigs. Only bit by bit was I able to write in 4 or 2. When my band would play out, entire sets would consist of instrumentals in 3. I remember from music theory classes that in early western notation used for chants in church singing, a circle indicated "Perfect" time, which was in 3. And Imperfect time (2 or 4 beats per measure) was a circle with a bite taken out, which looks like a C, and came to be called Common time. Probably someone can tell me I've got all this screwed up, but I found it intriguing...Now I can write in 4, I consider it progress of a sort!--Mer

[milfus]

ibanez, learn to juggleseriously, learn to juggle =0)

[ibanez468]

I can juggle! What's the point?

[gongchime]

Yep the Beatles knew all of that intuitively. In making this list I was thinking what would we have to ask a computer if we wanted to build a music composition program to compose good music. I've been looking at Expert Systems and the people analyzing what the experts are doing are not able to completely cover the task of asking the right questions in my opinion although they're getting better quickly. A lot of what humans are doing is inexplicable so after the fundamentals have been covered, maybe it's better to ask the experts what they would tell a computer to do instead of asking what the experts are doing so much. Experts are operating in the stratosphere and computers are operating in the dust. If the computer was a beginning student, would the expert start at the level of the stratosphere in his lessons to it? A better question might be, what parameters would an expert educator and musician give to a novice or intermediate student so that they don't get into trouble.Also, the list could make for an interesting party for musicians interested in random processes. Roll dice for each decision perhaps weighting the dice to prefer some choices over others and voila, tons of new, less crappy music than computer generated algorithm/fractal stuff. But such an endeavor is a huge bottle-neck. Unconsciously filtering out all of the bad possibilities is indeed much faster. I would really like a computer to just automatically run through all of the possibilities inherent in my starting motif or scanned in favorite tune and start playing them back one by one so I can listen for my favorite's then get to work recording them, instead of sweating over creating them.Mer perceptive of you to notice that the time signature was left out. How could I miss that? Crickey!3/4 time was perfect because it was coming from a Christian oriented background which sees heaven as prefect. Since they had a tripartite division of the heavens as in Father, Son and Holy Spirit, that was the reason they were using 3/4 for so much stuff in those days and in those cultures. Funny thing is, in Asia, 3/4 is a sign that the music still has it's Shamanic roots. I suspect the same is true of Pagan roots of Europe. Whereas music in 4/4 was a sign of the high culture of China as viewed by Korea, Japan and Vietnam. The comment coming from ethnomusicologists that Korean music retains it's shamanic roots and one clue is that it is in 3/4. Whereas many varieties of fairly traditional Japanese music don't use it. I just got back from Japan on a business trip to find a new market for my silver jewelry. I tried to listen to a Gagaku performance. The one still associated with the imperial court is the oldest continuously existing ensemble in the world at something like 1,400 years. But they only give performances once a year at the National Theater. There are other groups performaing at smaller shrines though. One was scheduled to play at a shrine for a funeral. The shrine had a website and had an English version. Clearly the Shinto priests were trying to get foreigners to attend but I'm not so sure the relatives of the deceased knew they were doing that. I chose not to attend but I did go to a shrine when there was no activity. They tie chords around doorways and gates to bind the evil spirits to keep them out. It didn't prevent me from entering so maybe their magic isn't working so well. HehehCurrently listening to the music of Mohamed El Bakkar associated with the oldest continuously existing dance tradition in the world. Belly Dance!

[milfus]

yeah but honestly, best case scenario, you make a computer that out does us all, and we all go bankrupt......

[ggalen]

As far as computers generating good music and putting us out of business, it could only happen for sectors where the music is simplistic and more purely "mathematical".But the music that is more "musical" and interesting...not to mention lyrics that tap into the deeper and more subtle human experience...ain't gonna happen anytime soon.As far as the simpler stuff, look at Garageband and loops: people move things around with the mouse until they get something they like.You can find an audience for some of the simpler stuff made that way, because people just flat-out like music, and there is an audience for most anything if you can find them.

[milfus]

really depends on the level of programming, i mean emotive music can be taught, so it stands to reason it could be programmed, just not easily

[gongchime]

Well, the music industry is looking at computers that listen to the radio and then have the computer try to compose music in the hit styles being played. Then those people won't have to pay writers. If big music companies are going to get this software, they're going to want to keep it out of other people's hands and have it all to themselves. But writers can still be independent and figure out ways to see cash rolling in. That might mean their own hit writing software bundle at home made by rogue computer programmers. Ever met one that wasn't? Me neither.Also computers can't go on tour yet so somebody has to play whatever is composed. They probably won't be advertising the fact that the music wasn't composed by the performer. I'm sure they'll try to get around having real people give concerts eventually too.

[milfus]

yeah, which helps musicians, composers however, would be obsolete completely, which is sad, although if the computers only listen to the radio im not really worrying about my job security just yet, hahanah I was just pointing out the fact that the ultimate goal is to put us all out of work, sometimes people need that perspective pointed out to them

[ggalen]

I built one of the first of these artificially intelligent "expert systems" 24 years ago. I did the expert system as my masters level thesis project at the University of Minnesota. It was used at the University of Minnesota Hospital to double check some physician interpretations of blood test results.And being a musician, I have often thought about this very topic, and how to write such a system.The thing is, in the early 80s they were predicting that intelligent computers would do all these things...yet they have never come close to doing most of them! Almost 25 years later!So I am more skeptical now.It just turns out to be a lot harder once you get in there and see what's really involved. That's been my experience in creating expert systems. And I have created a number of them once I graduated and got into corporate America.As far as the "hit analysis software", that computer program appeared about 3 years ago, and I have my doubts whether its pattern analysis of WAV files can predict a true hit better than an experienced producer whose brain has seen and analyzed all those patterns as well.The brain is an amazing "pattern-learning device". Look at 2-year-olds all learning language, just by listening to all this gobbledygook that they hear! I know they'd love to sell that software to record labels...except the record labels don't seem to have much money these days, eh?I always suspected somebody was just trying to get a bunch of venture capital money for their idea, but I really doubt that it went anywhere. If anyone knows more, I'd be interested to hear what happened.