electro-music.com   Dedicated to experimental electro-acoustic
and electronic music
 
    Front Page  |  Radio
 |  Media  |  Forum  |  Wiki  |  Links
Forum with support of Syndicator RSS
 FAQFAQ   CalendarCalendar   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   LinksLinks
 RegisterRegister   ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in  Chat RoomChat Room 
go to the radio page Live at electro-music.com radio 1 Please visit the chat
poster
 Forum index » Discussion » Composition
Bayesian Probability
Post new topic   Reply to topic Moderators: elektro80
Page 1 of 1 [14 Posts]
View unread posts
View new posts in the last week
Mark the topic unread :: View previous topic :: View next topic
Author Message
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Fri May 28, 2010 2:42 pm    Post subject: Bayesian Probability Reply with quote  Mark this post and the followings unread

Several years ago I read David Temperley's The Cognition of basic Musical Structures and was glad to see his extension of Lerdahl and Jackendoff's preference rules. In effect, this kind of rule incorporates informed intuition about musical context into analysis and provides substantial leverage in elucidating elements under the surface of musical score. On the down side it makes for a semi-formal system, one not amendable to translation into algorithms. In his new book, Music and Probability, Temperley shows how Bayesian Probability can be used to "statistically" formalize preference rules.

I know a number of people on this forum use probabilistic methods and was wondering if any are using Bayesian methods and if so, how and with what results?

_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Van_Tek



Joined: May 14, 2010
Posts: 3
Location: Kamloops, BC, Canada

PostPosted: Mon May 31, 2010 11:10 am    Post subject: Reply with quote  Mark this post and the followings unread

I've been producing music for 10 years now and I have no idea what you just said...
_________________
Dope Tito music: www.dopetito.com
My Facebook: www.facebook.com/VanTek
My Artist Page: www.facebook.com/jasonvantek
Back to top
View user's profile Send private message
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Mon May 31, 2010 5:52 pm    Post subject: Reply with quote  Mark this post and the followings unread

Van_Tek wrote:
I've been producing music for 10 years now and I have no idea what you just said...


Here are the three books that, if read and understood, would put one in a position to not only understand the initial post, but to assist with Project Zylaphon.

A Generative Theory of Tonal Music

The Cognition of Basic Musical Structures
and

Music and Probability

But, of course, I don't expect you to this--though I would rejoice at it. The purpose of the post is to engage someone who has read at least one of these books as a starting point for conversations on the topic of surface first analysis. I.E. starting with the notes provided by the composer and projecting deep structures from that data.

Sorry, it is a difficult area, and I need all the help I can find.
Embarassed Laughing

_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Fri Jun 04, 2010 2:49 pm    Post subject: Reply with quote  Mark this post and the followings unread

I wrote a patent application for a client in 2007 that made use of Bayesian data filtering to figure out the likelihood that you want to establish geographically local communication contacts as you travel, based on prior actions. (Simulations were quite scary in predicting human behavior. The patent is still pending.) I haven't done anything relating to music, although I did read Temperley's Music and Probability a few years ago.

The most useful newbie tutorial for me was Paul Graham's essays on Bayesian spam filtering,based on the statistical correlation of words within emails classified by the user as spam. It was a fairly understandable introductory essay.

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Sat Jun 05, 2010 5:13 am    Post subject: Reply with quote  Mark this post and the followings unread

Acoustic Interloper wrote:
... haven't done anything relating to music, although I did read Temperley's Music and Probability a few years ago.


Thanks, Do you remember what your impression was of Music and Probability?

_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Sun Jun 06, 2010 12:00 pm    Post subject: Reply with quote  Mark this post and the followings unread

bachus wrote:
Acoustic Interloper wrote:
... haven't done anything relating to music, although I did read Temperley's Music and Probability a few years ago.


Thanks, Do you remember what your impression was of Music and Probability?

I thought the book was fine, although tangentially related to my interests. As I recall it was about classification and generally related to score retrieval and categorization.

More to the point, I have general ideas about a representation format for scores derived from mechanisms in quantum mechanics, where events in a score have probabilities ranging from 0 to 1, and probabilities for events (both "atomic" events however they are defined, and their multi-layer hierarchical aggregates) are resolved to 0 or 1 only at performance time, with resolution determined in part by actual performance data coming into the score (from performers, conductor, room acoustics, etc.). Anything resolving to 1 plays and to 0 does not. The concept is inspired by the notion of measurement-as-participation and quantum entanglement. I am sure that I have prattled on about it before.

Even more to the point, I just got back from http://www.icmc2010.org/ this early AM and spent much of yesterday participating in the "Unconference," including a lengthy discussion among performers and some composers (ICMC has a strong classical bias; Unconference not as much) about the impossibility of repeat performances without the composer at hand, partly because the composer is now over-coupled into the job of instrument designer (designing unique, hard-to-duplicate instruments as a component of every score), partly because there is no portable, machine-and-performer readable score format capable of capturing scores that include strong, non-traditional sonic components.

From my perspective the notation should not be tool specific (C++, C#, Max, PD, Chuck, whatever). It must be a data language (e.g., on top of XML), readable by humans (via a custom music-atop-XML viewer), and extensible so that it can capture scores for unique instruments. XML is just an example, although a likely one. Clearly, I would like it to be able to handle probabilities, presumably Bayesian ones.

The other missing piece is the "technician" who nowadays is an additional performer who should get performance credit on the score and program. Someone referred to this person as the "Performer Whisperer."

I thought of your recent threads right away, and also how the Unconference discussion ought to be continued here. I did a bit of proselytizing for electro-music, with plans for more. I did also get some contacts for interested people, including one professor in my university system who teaches in middle PA. Following up on all this is in my to-do list.

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Mon Jun 07, 2010 6:41 am    Post subject: Reply with quote  Mark this post and the followings unread

Thanks Dale, interesting, informative post.

Acoustic Interloper wrote:
From my perspective the notation should not be tool specific (C++, C#, Max, PD, Chuck, whatever). It must be a data language (e.g., on top of XML), readable by humans (via a custom music-atop-XML viewer), and extensible so that it can capture scores for unique instruments. XML is just an example, although a likely one. Clearly, I would like it to be able to handle probabilities, presumably Bayesian ones.


Certainly notation must be extensible if it is to fill the needs of diverse approaches to music creation. But from my view point notation is most often a surface feature of an elaborate deep structure and to be functional in a computational sense, notation is inseparable from that structure. Though a data language like XML is certainly capable of encoding such deep structures it seems the whole shebang would have to be duplicated in some general-purpose language, e.g. Java, C# etc.. Which is a bit off-putting both conceptually and practically. That said, I certainly see the advantage of the XML/XML-Viewers model for surface feature presentation an interaction.

Perhaps I am missing something? Or perhaps that is the difference between viewing notation as a data to be interpreted by performers and viewing it as an interface into a deeper web of data to be manipulated by composers?

_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Mon Jun 07, 2010 2:38 pm    Post subject: Reply with quote  Mark this post and the followings unread

bachus wrote:
Or perhaps that is the difference between viewing notation as a data to be interpreted by performers and viewing it as an interface into a deeper web of data to be manipulated by composers?

These sorts of issues come up all the time in the design and implementation of CAD tools for VLSI design, which is where I spent much of my Bell Labs time (and where I met Mosc).

To the degree possible tools need to exchange data in a common, extensible format. In the "leaves" are the hooks for annotating information to tool-specific features, such as a pragma that indicates using some tool-specific VLSI compaction algorithm, for example. The core of an XML-hosted musical notation language would be standardized, representing common features. Some extensions might be expressed in terms of that core, e.g., aggregation of units defined in the core language. Some would be outside the language, i.e., resident in a specific tool or language. Such constructs should be in the "leaves" of a particular design file, as opposed to being over-coupled into the core from the outset. The XML, for example, could have URIs pointing to tool-specific annotations in a common, tool-neutral form, indicating precise ways to implement those common constructs in tool-specific ways.

Extension languages like Tcl or Python do this. Some extensions are in the language, built up using language constructs. Others are under-the-hood extensions under a foreign language API that appear as function calls in Tcl or Python. The Java Native Interface is similar, allowing access to underlying code written in C; that C code is the analog of the tool-specific features not found in the standardized language but nevertheless accessible from it.

In our discussion at the Unconference the performers expressed dissatisfaction with notations such as "depress pedal X for 2 beats" without any notion of what depressing pedal X represents sonically. The core concept such as "pink noise for 2 beats" should appear in the core of the score, with the means for realization (depressing a pedal in this example) "in the leaves." Sometimes the leaves are simply paths/URIs to machine-readable instructions.

Hope this makes sense. Part of my reading of the problem is that modern electro-classical composers over-couple too many things. They over-couple instrument design into the score, and they over-couple non-musical artifacts (from a performer's perspective) into the score as used at practice and performance time by a performer. The performers invest a lot of time preparing for a piece, and they do not want their efforts to be thrown away after 1 or 2 performances because of instruments that cannot be reconstituted, often for reasons of non-robust design, and because of scores that cannot be interpreted without the composer sitting in the room.

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Mon Jun 07, 2010 3:55 pm    Post subject: Reply with quote  Mark this post and the followings unread

Well, I certainly can be slow, but now I've got it and have to agree completely. Thanks for the clear and compelling explanation.
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Mon Jun 07, 2010 4:20 pm    Post subject: Reply with quote  Mark this post and the followings unread

Acoustic Interloper wrote:
The performers invest a lot of time preparing for a piece, and they do not want their efforts to be thrown away after 1 or 2 performances because of instruments that cannot be reconstituted, often for reasons of non-robust design, and because of scores that cannot be interpreted without the composer sitting in the room.


Actually I don't quite agree with every thing. Even though the creation of ephemeral compositions does not interest me, I have some misgivings about this. To my mind it is entirely valid for a composer to conceive and create a composition that is essentially ephemeral, having no life beyond a precious few performances other than recordings of those performances. So I would not want to disparage the preparatory efforts of performers in this and similar contexts as being "thrown away". But I guess that's a composer's view. I'm sure if I were a performer I would piss and moan about it.

_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Mon Jun 07, 2010 4:52 pm    Post subject: Reply with quote  Mark this post and the followings unread

bachus wrote:
Acoustic Interloper wrote:
The performers invest a lot of time preparing for a piece, and they do not want their efforts to be thrown away after 1 or 2 performances because of instruments that cannot be reconstituted, often for reasons of non-robust design, and because of scores that cannot be interpreted without the composer sitting in the room.


Actually I don't quite agree with every thing. Even though the creation of ephemeral compositions does not interest me, I have some misgivings about this. To my mind it is entirely valid for a composer to conceive and create a composition that is essentially ephemeral, having no life beyond a precious few performances other than recordings of those performances. So I would not want to disparage the preparatory efforts of performers in this and similar contexts as being "thrown away". But I guess that's a composer's view. I'm sure if I were a performer I would piss and moan about it.

I agree with you, and I think the performers in the room would have agreed, that it should be *possible* to write and perform ephemeral compositions. Their main point was that the composition should not be ephemeral simply because they as performers cannot repeat a performance of a heavily practiced and otherwise non-ephemeral piece simply because the communication of that piece to them is incomplete, requiring the composer to be lurking in the shadows pushing buttons during performance.

Of course, I play improvised music and none of this much affects me directly. After doing *Sketches of Spain* with classical performers who had no inclination towards improvisation, Miles Davis referred to that mode of performance as "robot shit." Improvisation in Indian classical music also came up during this discussion.

But I am, after all, a jazzified folkie and generally a musical poser. Twisted Evil

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
bachus



Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5

PostPosted: Tue Jun 08, 2010 5:29 am    Post subject: Reply with quote  Mark this post and the followings unread

OK. I confess, the last two lines were a joke meant to poke fun at the attitudes of both composers and performers. -- After all I am a would-be composer Laughing
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Tue Jun 08, 2010 7:55 am    Post subject: Reply with quote  Mark this post and the followings unread

bachus wrote:
OK. I confess, the last two lines were a joke meant to poke fun at the attitudes of both composers and performers. -- After all I am a would-be composer Laughing

Oh, there was plenty of implied pissing and moaning. Laughing

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
Acoustic Interloper



Joined: Jul 07, 2007
Posts: 2067
Location: Berks County, PA
Audio files: 89

PostPosted: Sun Aug 21, 2011 8:03 pm    Post subject: Reply with quote  Mark this post and the followings unread

Acoustic Interloper wrote:

Even more to the point, I just got back from http://www.icmc2010.org/ this early AM and spent much of yesterday participating in the "Unconference," including a lengthy discussion among performers and some composers (ICMC has a strong classical bias; Unconference not as much) about the impossibility of repeat performances without the composer at hand, partly because the composer is now over-coupled into the job of instrument designer (designing unique, hard-to-duplicate instruments as a component of every score), partly because there is no portable, machine-and-performer readable score format capable of capturing scores that include strong, non-traditional sonic components.

I just ran across a transcript of this Unconference here.

_________________
When the stream is deep
my wild little dog frolics,
when shallow, she drinks.
Back to top
View user's profile Send private message Visit poster's website
Display posts from previous:   
Post new topic   Reply to topic Moderators: elektro80
Page 1 of 1 [14 Posts]
View unread posts
View new posts in the last week
Mark the topic unread :: View previous topic :: View next topic
 Forum index » Discussion » Composition
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum


Forum with support of Syndicator RSS
Powered by phpBB © 2001, 2005 phpBB Group
Copyright © 2003 through 2009 by electro-music.com - Conditions Of Use