-
Better program it to do GB gigs as well. Electricity isn't free...
-
12-04-2017 10:57 AM
-
There's a paper floating around where a researcher uses Markov chains to analyze Coltrane transcriptions:
https://theses.lib.vt.edu/theses/ava...ted/dmfetd.pdf
Frankly, I think it would be very difficult to create a Chomsky-like system for improvisation, if only because (as I understand it) Chomsky's idea is that the "universal grammar" underlying all human-languages is a reflection of the actual base structures and processes of human thought and perception. But there's no real correlation for jazz improvisation - it's a cultural practice specific to a particular time and place, and a relatively recent one at that.
If I were forced to take a stab at creating an improvising computer, I would take more of a machine-learning approach, and then feed it transcriptions from as many jazz solos as I could until it "kind-of-sort-of-got-it." I predict the best it could do is create decidedly boring solos filled with cliches, but who knows, it might surprise me.
-
OP: "I am trying to do research about the learnability of jazz technique as a RULE-BASED SYSTEM, potentially leaning towards deriving a formal grammar of a given instruction system in terms of substitution rules."
Christian: "Improvisation is never based on rules."
OP again: "Agreed. But, could I please ask the fellow forum members to not imply things I did not say or write?"
You might not think you IMPLIED it, but you made it easy for us to INFER that you want to make instructions for removing the soul from jazz, which is probably why Christian gave you the responses he did.
-
Originally Posted by Pat Clare
-
Yeah I’ll say it again. There seems a confusion here between practice and improvisation.
When we practice we are like machines. When we play we seek to know God (insert whatever terminology that can refer to the Great Mystery if uncomfortable with the G word.)
Plugging the rules into a machine should be straightforward. I would argue that as mathematicians regard a problem soluble by a computer as ‘not a good problem’, we should regard these aspects of music for what they are - the tools of the trade. Nothing more.
-
Improvising trombonist/composer/thinker George Lewis has long engaged with computers and improvisation.
Here's a link to a lecture he presented on the subject.
-
Originally Posted by zirenius
-
Originally Posted by matt.guitarteacher
-
Originally Posted by christianm77
I think there are more rules to learn than those governing language, but this might be interesting to some:
-
Originally Posted by christianm77
Also the are NO rules that is the whole idea of real Jazz. The only judge is your ears and like talking to people not everyone is going to like or agree with what they hear, but if you are being true to yourself and speak or play with conviction then all is good.
-
Originally Posted by destinytot
Flippancy aside, I might quote Peter Bernstein (c/o Jonah) - 'a machine can play the right notes.'
-
All joking aside, there are a lot of educated jazz guitar instructors on YT whose playing sounds like a structured algorithm on a computer. They lack soul, but they may be able to help you.
-
Originally Posted by eh6794-2.0
-
A book that might be useful is Jerry Coker's "Elements of the jazz language for the developing inprovisor". He gives 20 devices used in improvised jazz, and claims that just about all solos on chord progressions can be analysed with them
-
Originally Posted by eh6794-2.0
Even practice shouldn't sound mechanical, should sound like music. When learning some new lets say a scale when learning the fingering then being mechanical is okay, but once you have it under the fingers then time to just focus on making it sound musical even when practicing. What you practice is what will come thru in your playing so do you want to sound like scale exercises or sound like music with feel, that has been said my many great musicians.
-
Ok, wow, I just had to work a couple of hours and this thread grew enormously. Thank you all and consider this a placeholder, as I will have to reply adequatly after preparing dinner (or I might place the very much more substantial problem of my wife being unhappy about the lack thereof). Be right back.
-
Uncle Willie's 'Jazz Everyone'?
-
Originally Posted by docbop
-
Not to be snarky, but why get a degree in "systematic musicology" ?
My college age son had a story about an English Dept. "get to know us" reception, and he was greatly relieved when another kid stated, out loud, what many there were afraid to say, namely "They liked to read good books...but hated talking about them."
-
Originally Posted by bako
George Lewis is a genius, a living legend, and has been exploring the intersection between improvisation and computers for decades now. I saw an amazing concert with Roscoe Mitchell and Muhal Richard Abrams (RIP) where he used some of those techniques.
He's certainly not the only one. Just the other day, I saw Miles Okazaki using a system he created that interacted with his improvisations.
-
Originally Posted by dasein
-
I think it’s an interesting topic. I’d have a crack at it myself if my programming skills were up to it, would be a fun project.
Might be worth checking out the Wayne Krantz OS for Improvisers, I don’t have it but it sounds like it might be of interest.
Also Jerry Bergonzi uses simple rules and patterns to generate new ideas. E.g. play 3, 4 or 5 note shapes with a simple rule such as the notes must go up, up, down etc. then take the resulting pattern all the way through the changes of a tune.
-
[QUOTE=dasein;823339]It's always funny when people on this forum -- who, let's be honest, are mostly hobbyists and amateurs -- declare with great authority what is and what isn't permissable, when professionals and accomplished artists have already been happily exploring it for years.
I'm not saying I'm in favor of passing a law prohibiting this type of thing. But of all the music I like, and that I've ever liked, I'm pretty sure none of it was created this way---algorithms, etc.
I just think that attempting to reduce art to a formula, or algorithm, is barking up the wrong tree. But if someone wants to spend their life doing that, I can't stop them.
But I don't have to like it.
Not hard to fathom.
Is it??
Really.
-
Originally Posted by dasein
-
First off I am a pretty poor musician, but maybe I can help. When I first started to learn guitar and about music in general, I hoped I could play to my strengths which are engineering simulation and programming, to come up with an algorithm that would allow me to play jazz. I now realise that was a stupid approach. However I did find a few books that might be of interest that could probably be used for programming some cod jazz. As already mentioned the David Baker books called "How to play Bebop" are generally a load of lines over stock progressions that might be of use, but one that cones to mind that has a small number of "rules" or at least possibilities that could be played with, is Jerry Coker's "Elements of the Jazz Language for the Developing Improviser" which I think is a very good book, as it often reminds me what it is I have just heard. Also this academic has an approach that may be useful.
Reno De Stefano, jazz guitarist
I realise now that in the few minutes it took to write this response, and then forgetting to press send immediately, everything I have written has been said by someone else!Last edited by plasticpigeon; 12-04-2017 at 05:32 PM.
Henriksen Blu 6 w/ gig bag
Today, 03:29 PM in For Sale