A group of tones can have many chord names It;s a contextual thing.. It depend's on what key you are in you can have a missing root you can have other missing tones but that would affect the name of the chord. but beyond that the same group of tones (or their inversion) can be intepreted many ways For instance a C chord (C-E G )could be a considered an Em #5 ( E -G B#) or, a bit of a stretch, as a G sus4 aug 5 No third It's contextual. take a harmony course On Sat, 14 Jun 2003 04:17:11 GMT, "Angeline J." <email@example.com> wrote: > >"Gerry Scott-Moore" <firstname.lastname@example.org> wrote in message news:email@example.com... >> In article <gqoGa.5193$Jw6.firstname.lastname@example.org>, Angeline >> J. <email@example.com> wrote: >> >> > What kind of weird chords do you believe a computer can't be >> > taught to identify? >> > >> > George Shearing's "Over the Rainbow" (his arrangement) >> > >> > In a good portion of the chord progressions, the root note is left >> > out. >> >> Leaving the root note out of the chord doesn't automatically produce an >> inaccurately identified chord. Clearly a program doesn't depend on >> analyzing chords only if the root is present. >> >> Are you gonna spell me a chord that you think a computer can't analyze >> or not? I'll assume you won't attempt it. > >I'll break out the chord progression from "Over the Rainbow" later. > > >> > I'm not sure what computer algorithm could be devised to recognize >> > that. >> >> Do you know a lot about computer algorithms? > > >Yes. One of my majors in college was Computer Science. I went through the Math Department at UC San Diego and much of the upper division coursework was theoretical although we did do a fair amount of real programming. > >Programs like compilers (translate third generation languages like C, BASIC, etc. into assembly code) parse a program, break it up into different symbols and run it through an algorithm which (IIRC) models a structure called a deterministic finite automat a. It works for computer languages because the grammar is fixed. A deterministic finite automata does not work for spoken languages - there is too much variation. > >Chord recognition is almost infinitely simpler than understand a spoken language, but it still isn't deterministic because of context (for example, the "implied harmony"). However, I'll guess that an algorithm could be designed that would recognize all bu t the most esoteric jazz progressions.