-
Dear folks of this wonderful forum,
I have an enquiry considering my Master's thesis in systematic musicology.
The subject of my research is titled "Jazz as a specific field of computer-based modelling".
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. (For the linguistically inclined: the ideal aim would be a context-free or "mildly context-sensitive" grammar on the Chomsky hierarchy)
So far I have implemented Barry Harris' approach for learning descending scalar material with added chromaticism in a Python script that can now "improvise" (aka randomly pick) Jazz lines based on a rule system. I need and want to expand this.
Therefore, I am looking for other jazz instruction material that focusses on a limited set of rules/devices that are modified and could be regarded as a rule-based system. Please note that I am not implying that improvisation can be reduced to rules, but I am intrigued by the possibility of using rule-based systems to further my understanding of technical instruction in the field of jazz guitar.
So far, I have been looking at Randy Vincent's work (cellular system) and Ligon's books. Are there any other systems coming to your minds that might be worth looking into to expand my little experiment? I have been talking to the man behind Band in a Box, but I don't think that people from the music software industry would be willing to share too much of their source codes (which I can perfectly understand). So, can any of you help a student with his research? Obviously, I could share the results of what I am doing - even though I am convinced that the software results would be of marginal interest to the professionals here (as in, they are musically still very basic).
Have a great day,
Sebastian
-
12-04-2017 03:37 AM
-
Im not a Luddite normally
Elon Musk is my hero ....
But I think this is a bad idea ...
At best you'll create a machine that performs a clever
party trick ... the output won't have meaning ...
Sorry to be a wet blanket
-
No offences taken. However, the entire point is not creating a machine that can do any kind of trick. Also, I do not want to create something that I expect to play at a musically convincing level. I don't even so much care about any specific quality of the output. I am interested in mechanisms behind the acquisition of improvisational skills and musical material needed to improvise.
Originally Posted by pingu
The idea is the following:
(1): Technical instruction on how to improvise is a useful way to acquire the necessary skills to do so.
(2): Some technical instruction is organised in a hierarchical fashion that resembles a rule-based system.
If (2) is true, then the technical instruction can be translated into something that a certain kind of computer can execute.
I am merely hoping to gain some insight into how improvisational systems can be conveyed when the personal and manual level of having to learn the instrument (and other external influences) are left out of the equation. Certainly, there is no harm to be done here and the outcome might be interesting - especially considering that my interest is also that of a guitar teacher, who might learn a thing or two about the concepts and their inner workings that we usually don't think about when practicing and / or teaching them, right?
Distinctly: I think that being able to reformulate the implications of a given system for studying jazz improvisation in terms that even a computer could understand and execute is in itself already a potentially helpful thing in teaching.
Trust me, the current state of "improvisation" in my program is nothing to write home about - it is nowhere near even hoping to become a party trick one day.
But yes, thank you for your warning. I don't mean to convince anybody of the usefulness of this idea - I would, however, very much appreciate any hints at instruction material that might be productive here. Thank you for your time.
-
Improvisation is never based on rules.
Practice is based on rules.
-
By all means: Originally Posted by christianm77
Agreed. But, could I please ask the fellow forum members to not imply things I did not say or write?
in my first post, I stressed the fact that improvisation cannot be reduced to rules. In my second post, I tried to be more specific.
To be very clear about my actual question: are you aware of any instructional material for the technical side of jazz improvisation that in your opinion gives clear rules for what is to be practiced?
Please note:
1. if you are afraid that this is aiming at diminuishing the role of musicians: it is not, why should I harm my own profession?
2. if the questions above should not be discussed here - fair enough.
I would just like to ask anybody informing me about the impossibility to reduce improvisation to computable rules to note that I am aware of the fact and there is no need in pointing it out to me.
BTW: if 2 holds true, I will be glad to remove my posts and stick to less dubious topics.
-
A computer cannot improvise because a computer cannot know God. Originally Posted by zirenius
By all means:
Agreed. But, could I please ask the fellow forum members to not imply things I did not say or write?
in my first post, I stressed the fact that improvisation cannot be reduced to rules. In my second post, I tried to be more specific.
To be very clear about my actual question: are you aware of any instructional material for the technical side of jazz improvisation that in your opinion gives clear rules for what is to be practiced?
Please note:
1. if you are afraid that this is aiming at diminuishing the role of musicians: it is not, why should I harm my own profession?
2. if the questions above should not be discussed here - fair enough.
I would just like to ask anybody informing me about the impossibility to reduce improvisation to computable rules to note that I am aware of the fact and there is no need in pointing it out to me.
BTW: if 2 holds true, I will be glad to remove my posts and stick to less dubious topics.
-
First of all, I just checked your website. You seem to be a very accomplished musician and. I am grateful that you are taking the time to reply here. Originally Posted by christianm77
back on topic:
1. Please don‘t plainly ignore the text I type, it‘s not nice - no matter, what point your musical carreer is at.
2. No need to argue: I never claimed that a computer can improvise. I want to do research about how the technical and conceptual skills neccessary are acquired and could be described in a specific way (a computational model).
3. I did not mean for this to have any spiritual or religious dimension. No harm meant, but are you sure that you meant to imply that non-religious beings cannot improvise? I am taking no offense, my other major being theology, but I am equally sure that other prominent colleagues might disagree. I value the spiritual dimension of music very dearly, but using it to shut down academic discussion of certain aspects of our traits seems like dangerous territory to me...
-
Sorry I am a bit guilty of not really reading the OP... And being a bit of an internet troublemaker to boot.
I think the whole issue of getting machines to improvise raises interesting philosophical and politico-economic issues, and the path of the musician is fundamentally different to the path of the scientist. I know, because I have been both in my life.
Jazz technique, encapsulated here by the rules of Barry Harris, can be completely automated. I'm pretty sure I could write a spreadsheet (my coding skills are limited - I'm so old they taught me FORTRAN at uni) that would construct lines using the BH framework today if I wanted to. Or (as I prefer) you could use dice and a table to randomise your options (I like the tactile aspect.)
BH is the system I am most familiar with. I can't advise on the others.
The question is, though, why? What do you think is the benefit to the human improviser?
-
How do we know you're not a computer model of a grad student asking other people to do his/her research for him/her? Originally Posted by zirenius
Sent from my SM-J700T using Tapatalk
-
public class HelloWorld { public static void main(String[] args) { // Prints "Hello, God" in the terminal window. System.out.println("Hello, God"); } } Originally Posted by christianm77
Sez who?
Sent from my SM-J700T using TapatalkLast edited by John A.; 12-04-2017 at 11:30 AM.
-
Man, keep it simple for this dinosaur, no OOP, pls. Originally Posted by John A.
10 Print "Hello, God"
20 Goto 10
-
This raises the interesting spectre of a jazz Turing test.
The interesting thing about all of this is that we are looking at it from an external sense can a human tell the difference between a computer or a human playing jazz (of course for true AI, we would need to reverse it - can a computer spot a computer?)
But the musician's path is completely subjective. Objectivity is impossible.
-
@John A. (1st post): you don't. what difference does it make?
@John A. (2nd post): appreciate the humour.
@christianm77: appreciate your last post. Thank you for that. Brief answer two your "why question":
Because this field of study might actually be a relevant topic within the musicological side of cognitive science(s), more specifically in looking at areas of shared interest for computational linguistics and musicology.
Good on you for being both a scientist and an artist in your life. So far I have truly achieved neither and am working on both.
---
Just saw your last post while typing. Both implied questions are - despite intended humour - actually quite relevant to the topic at hand, but very much outside the scope of the current question. I am also not even claiming any kind of something resembling intelligence here. Geez, I thought that I phrased my enquiry quite carefully...
-
If you wanted to come at it from another angle, I reckon you could analyse and catalogue Coltrane's lines in his Giant Steps solo against the relevant chords they occur on, and formulate some rules from it. There are quite a lot of repeated patterns in that solo.
-
And here we have encapsulated the difference between Music Theory proper (study of music) as opposed to what we think of as music theory (study of how to make music) Originally Posted by zirenius
@John A. (1st post): you don't. what difference does it make?
@John A. (2nd post): appreciate the humour.
@christianm77: appreciate your last post. Thank you for that. Brief answer two your "why question":
Because this field of study might actually be a relevant topic within the musicological side of cognitive science(s), more specifically in looking at areas of shared interest for computational linguistics and musicology.
Good on you for being both a scientist and an artist in your life. So far I have truly achieved neither and am working on both.
---
Just saw your last post while typing. Both implied questions are - despite intended humour - actually quite relevant to the topic at hand, but very much outside the scope of the current question. I am also not even claiming any kind of something resembling intelligence here. Geez, I thought that I phrased my enquiry quite carefully...
You did phrase your question carefully, but unfortunately I did not read it carefully.
As a professional musician I get very touchy about things like this. People should understand our work is being automated out of existence and society shows no sign of catching up.
Research carried out for blue skies purposes does not always remain so.
-
Tbh the way a musician practices is an algorithm. We will try and apply the same material consistently through a tune. Triads, 1 3 5, scales etc.
I think the area of interest is not what we practice, but what happens in performance in the moment of improvisation.
Which is where we can find ourselves playing something we haven’t played before.
Also, practice for high level jazz musicians is as concerned with flexibility rather than ingraining patterns.
-
Originally Posted by christianm77
-
@Christian: Thank you very much. Don't worry, I can perfectly understand. I am mainly a musician and guitar teacher myself. I am branching out into academic enquiries, because I am really interested in grasping the mechanisms behind what I like doing best and I don't want to have to rely on just music financially for the rest of my life. Simple as that...
Honestly speaking, whatever comes out of this will not be able to replace a musician in any way. This entire research is not even glancing at inter-musician interaction, for example. As of now, it is not even taking accompaniment into account. It merely generates 4-part chords and randomly generates lines on top that are derived from a limited set of rules. So yes, nothing to fear here and no evil master plan in my head...
-
agreed. also: light years ahead of what technology can capture as far as I know. And I am quite happy that it's this way. Originally Posted by christianm77
-
I think this is very interesting, and I'll try to make a small (on-topic) contribution later in the week. Originally Posted by zirenius
@John A. (1st post): you don't. what difference does it make?
@John A. (2nd post): appreciate the humour.
@christianm77: appreciate your last post. Thank you for that. Brief answer two your "why question":
Because this field of study might actually be a relevant topic within the musicological side of cognitive science(s), more specifically in looking at areas of shared interest for computational linguistics and musicology.
Good on you for being both a scientist and an artist in your life. So far I have truly achieved neither and am working on both.
---
Just saw your last post while typing. Both implied questions are - despite intended humour - actually quite relevant to the topic at hand, but very much outside the scope of the current question. I am also not even claiming any kind of something resembling intelligence here. Geez, I thought that I phrased my enquiry quite carefully...
For now, though, I'll just express the opinion that autonomous learners are the ones to watch in 'jazz'.
I don't think much can be added to Clark Terry's "Imitate, assimilate, innovate" - except perhaps "Don't listen to pinheads who can't play worth shit" - (though that might frighten the horses).
-
It's seems you should be able to create some organized controlled choices of improv. But without understanding the organization behind the improv. your going to miss many of the possible choices of the performance.
So first you need to understand the the interacting parts of the performance. (you need to understand).
Be aware of how musicians react to using the understanding,(the theory and common practice).
Be able to rate the possible choices of what to play in the performance...weight of history, weight of interaction in context, external influences, the audience etc...
You could limit the output to one specific player or method...
Sounds interesting, good luck.
-
The other name that comes immediately to my mind is David Baker.
He played jazz, composed tunes, transcribed / analyzed many of the greats (Clifford Brown, Miles, Coltrane, and Sonny Rollins), taught at the university level, and wrote extensively about jazz improvisation and pedagogy.
Here's a video of him talking to a group of schoolchildren about taking up the trombone in 7th grade and being told by a teacher he had no talent. From there, he went on to make his own tuba.... Interesting story of how one musician developed.
Here is the material of his that can be bought from Jamey Aebersold's jazzbooks.com site. "How to Play Bebop" might be his most influential work (in 3 volumes).
jazzbooks.com: Search Results
Good luck with your research.
-
He lives! Good to see you posting. :-) Originally Posted by MarkRhodes
-
Will the machine have to make a living to support itself, or will it be given a free ride?
-
Patterns and lines can probably be programmed. I think Connecting Chords with Linear Harmony is a good choice, as you said that you are already referencing Ligon's material.
However, the notion of improvised motif's and their subsequent variation/development would seem to pose an additional problem.
Reply With Quote
Has anyone played or had a Supro Amulet ?
Today, 04:44 AM in Guitar, Amps & Gizmos