Muzza

Hi guys. I've got a patch set up in a way that it stores the midi note numbers being played in seperate int boxes (eg. playing a c Major triad stores 48 52 55 independantly). How do i get Max to analyse these numbers to find the root of the chord, baring in mind the chords could be in any position (not always root) and any type (major, minor, dim 7th) ?

analysing-chords

Sounds like a job for Peter Elsea's fuzzy logic stuff. He has a tutorial

that talks about exactly this kind of thing.

> Hi guys. I've got a patch set up in a way that it stores the midi note numbers

> being played in seperate int boxes (eg. playing a c Major triad stores 48 52

> 55 independantly). How do i get Max to analyse these numbers to find the root

> of the chord, baring in mind the chords could be in any position (not always

> root) and any type (major, minor, dim 7th) ?

Sounds like a job for Peter Elsea's fuzzy logic stuff.  He has a tutorial
that talks about exactly this kind of thing.

On 3/19/07 8:04 PM, "Bill Murray"  wrote:

>
> Hi guys. I've got a patch set up in a way that it stores the midi note numbers
> being played in seperate int boxes (eg. playing a c Major triad stores 48 52
> 55 independantly). How do i get Max to analyse these numbers to find the root
> of the chord, baring in mind the chords could be in any position (not always
> root) and any type (major, minor, dim 7th) ?

Cheers
Gary Lee Nelson
Oberlin College
www.timara.oberlin.edu/GaryLeeNelson

For a further look at this, you should check out Robert Rowe's books.

On Mar 19, 2007, at 8:04 PM, Bill Murray wrote:

> Hi guys. I've got a patch set up in a way that it stores the midi note 

> numbers being played in seperate int boxes (eg. playing a c Major 

> triad stores 48 52 55 independantly). How do i get Max to analyse 

> these numbers to find the root of the chord, baring in mind the chords 

> could be in any position (not always root) and any type (major, minor, 

>
> Hi guys. I've got a patch set up in a way that it stores the midi note 
> numbers being played in seperate int boxes (eg. playing a c Major 
> triad stores 48 52 55 independantly). How do i get Max to analyse 
> these numbers to find the root of the chord, baring in mind the chords 
> could be in any position (not always root) and any type (major, minor, 
> dim 7th) ?
>
>
www.petermcculloch.com


by comparing your three numbers versus a predefined set

of chords (which are also lists of three numbers) using +,-,

by comparing your three numbers versus a predefined set
of chords (which are also lists of three numbers) using +,-,
pak, unpack, ==, zl, ...

Check out the books 'Machine Musicianship' by Robert Rowe and 

'Composing Interactive Music' by Todd Winkler, both published by MIT 

Press - both have good explanations of various algorithms to 

accomplish these types of tasks, and working example Max patches on 

On Mar 19, 2007, at 6:04 PM, Bill Murray wrote:

> Hi guys. I've got a patch set up in a way that it stores the midi 

> note numbers being played in seperate int boxes (eg. playing a c 

> Major triad stores 48 52 55 independantly). How do i get Max to 

> analyse these numbers to find the root of the chord, baring in mind 

> the chords could be in any position (not always root) and any type 

Check out the books 'Machine Musicianship' by Robert Rowe and  
'Composing Interactive Music' by Todd Winkler, both published by MIT  
Press - both have good explanations of various algorithms to  
accomplish these types of tasks, and working example Max patches on  
the accompanying CD-ROMs.

>
> Hi guys. I've got a patch set up in a way that it stores the midi  
> note numbers being played in seperate int boxes (eg. playing a c  
> Major triad stores 48 52 55 independantly). How do i get Max to  
> analyse these numbers to find the root of the chord, baring in mind  
> the chords could be in any position (not always root) and any type  
> (major, minor, dim 7th) ?

Think about using twelve-tone theory, in that you want to distill any combination / position etc. to its prime form... so a major chord is 0 4 7, a minor one is 0 3 7, etc..... Would take some doing but you could then compare any combination to a predetermined list of chords of any level of complexity.

Elsea's tutorial, and I am a bit fuzzy on this, shows chords represented as

a binary vector where each of twelve values represents the presence or

absence of a pitch class. For example, C major (0 4 7) would be

If you rotate F major 5 positions left it matches C major. This method can

be used to identify chord types. Rotate until you get a match. The number

On 3/19/07 11:41 PM, "Seejay James" wrote:

> Think about using twelve-tone theory, in that you want to distill any

> combination / position etc. to its prime form... so a major chord is 0 4 7, a

> minor one is 0 3 7, etc..... Would take some doing but you could then compare

> any combination to a predetermined list of chords of any level of complexity.

Elsea's tutorial, and I am a bit fuzzy on this, shows chords represented as
a binary vector where each of twelve values represents the presence or
absence of a pitch class. For example, C major (0 4 7) would be

If you rotate F major 5 positions left it matches C major.  This method can
be used to identify chord types.  Rotate until you get a match.  The number
of rotations (mod 12) is the root.

On 3/19/07 11:41 PM, "Seejay James"  wrote:

>
> Think about using twelve-tone theory, in that you want to distill any
> combination / position etc. to its prime form... so a major chord is 0 4 7, a
> minor one is 0 3 7, etc..... Would take some doing but you could then compare
> any combination to a predetermined list of chords of any level of complexity.
>
> --CJ
>

> Check out the books 'Machine Musicianship' by Robert Rowe and 'Composing 

> Interactive Music' by Todd Winkler, both published by MIT Press - both 

> have good explanations of various algorithms to accomplish these types 

> of tasks, and working example Max patches on the accompanying CD-ROMs.

a [at.chord] external that detects almost any chord there is. And there 

is [at.tonic] which tells you the tonic you're playing in.

Drawback is that it requires FTM 2.0 and Iracm still hasen't realease 

the final version of it... (and there is no Windows version right now). 

At least on OS X you could try to compile FTM 2.0 yourselfe if you have 

Steven Miller wrote:
> Check out the books 'Machine Musicianship' by Robert Rowe and 'Composing 
> Interactive Music' by Todd Winkler, both published by MIT Press - both 
> have good explanations of various algorithms to accomplish these types 
> of tasks, and working example Max patches on the accompanying CD-ROMs.

Or check this one: http://artificial-tango.nullmedium.de/ - it includes 
a [at.chord] external that detects almost any chord there is. And there 
is [at.tonic] which tells you the tonic you're playing in.
Drawback is that it requires FTM 2.0 and Iracm still hasen't realease 
the final version of it... (and there is no Windows version right now). 
At least on OS X you could try to compile FTM 2.0 yourselfe if you have 
a day to spare.

Analysing Chords