 sj |
||
|
||
systematic tune and song structure?
|
|
|
|||||||
|
systematic tune and song structure? |
Share Thread
|
||||||
|
Subject: Comparison of songs, tunes etc From: pavane Date: 02 Aug 01 - 10:48 AM Does anyone know if there is a systematic method already defined to classify song (lyric) and tune structures? I learned about such things as Iambic Pentameters at school, but that is only part of the story. We should be able to describe a song as having verses with n lines, the structure of each line (where lines may be different) of, say n syllables, the accenting being on which of them, i.e. on every other, like 2/4 time, every third, like 3/4 or 6/8, and so on. If such a method were in use, we could more easily compare songs, tunes, and match songs to tunes.
Note: I do know that there are various methods to extract common features from tunes, for matching purposes, the simplest being the 'Up/Down/Same' method, where each note is given one of the three codes relative to the previous note. In this format, Villikins & Dinah would start
|
|
Subject: RE: systematic tune and song structure? From: SharonA Date: 02 Aug 01 - 10:57 AM If there isn't such a system, what a life's work it would be to create one! |
|
Subject: RE: systematic tune and song structure? From: MMario Date: 02 Aug 01 - 11:16 AM Talk to Bruce Olsen - |
|
Subject: RE: systematic tune and song structure? From: Burke Date: 02 Aug 01 - 11:45 AM For English songs written in stanzas there are "meters" that describe the number of syllables in the lines & how many lines. Since the commonest number of lines is 4 that is not always stated. You can see it used most often in hymn books & religious tune books. The poetry is marked with the meter & the tunes list their meter. Hymnals frequently have meter indexes. Heres the one for the Sacred Harp. Here's the one for Cyberhymnal
The most frequently used meters are named:
Less used are just listed by syllables. It can get more complicated when choruses are added. The chorus can be of a different meter from the basic text. Amazing Grace, Gilligan's Island & loads of other English poetry are CM, that's why you can swap the tunes & why it's called Common meter.
|
|
Subject: RE: systematic tune and song structure? From: M.Ted Date: 02 Aug 01 - 11:50 AM The short answer is yes. Or no, depending on what you mean. A lot of different people have analyzed folk music, and or folk songs, but they use differing methods and differing analytical tools, depending on what they are looking for. It is certainly easy to analyze the lyrical structure of the verse, if that is what you want to do, using the names like dactylic hexameter, and to number the lines, and using names for line/verse structure that are used in poetic analysis, like Petrarchan or Quantrain, or whatever else-- Different verse structures and patterns of accent have different names, often based on the poet who used them, or the time and place that they were popular-- There is also a language for the structural analysis of music--since folk music generally is musically fairly straightforward, structurally (it tends to follows the verse structure) and, as has been pointed out many times here, there are many songs but relatively few melodies, people who analyze the music are usually more interested in the histories, evolution,derivations and transmission processes, than of the purely structural aspects of the music. How you analyze things really depends on what your objective is--the basic analyses correspond to the basic elements, which are scale and meter for melody and the rules of verse for the lyrics-
|
|
Subject: RE: systematic tune and song structure? From: pavane Date: 02 Aug 01 - 11:52 AM Yes, I have seen such things, but they are intended for people to read. Ballad Stanza is one I have seen which is used for songs. I am looking for a system in which it would be easy for a computer program to make comparisons. I will have a think about using something like XML for it. |
|
Subject: RE: systematic tune and song structure? From: GeorgeH Date: 02 Aug 01 - 11:57 AM But why should you want to?? Part of the beauty of Folk Song is that whatever structure you attempt to ascribe to it you'll so often find that it's the exceptions to the rule which give the thing its beauty . . Or as Martin Simpson was sometimes heard to say, when introducing Betsy the Serving Maid "If anyone wants to work out what time signature this is in - good luck!" (And that was when he sang it slowly . . .) G. |
|
Subject: RE: systematic tune and song structure? From: M.Ted Date: 02 Aug 01 - 01:38 PM But what is it you want to compare? If it is meter, remember that meter depends of syllabic stress, which is not apparent from word spellings. You'd have to develop a way to enter words that indicated where the syllable were, and whether they were stressed or unstressed--basically, you'd have to figure out the meter of a lyric before you could enter it-- As to melody, when you notate a piece of music, you create a representation of it's rhythmic an melodic structure that then is used to recreate the piece. Again, what do you want to compare? Why do you need a computer to do it? |
|
Subject: RE: systematic tune and song structure? From: McGrath of Harlow Date: 02 Aug 01 - 02:00 PM When the song collectors were going round at the start of the last century, they'd find that singers would break all the rules. They'd change the metre, they'd sing notes that weren't in the appropriate key, they'd vary the tune from stanza to stanza, they'd vary the length of the stanzas.
So a lot of the time they'd treat these as mistakes to be corrected, and see themselves as "restoring" the song to the way it was supposed to be. (And maybe clean up the verses while they were about it.)
And that was even without having computers. I suppose now they could rectify the field recordings so that the singers could hear what it was they were supposed to be doing. And then do it again, but this time do it right... |
|
Subject: RE: systematic tune and song structure? From: pavane Date: 02 Aug 01 - 02:30 PM No, the intention is not to 'correct' or to alter anything, just to make comparisons easier. And comparison of tunes can be important (and lucrative) where copyright infringements are claimed. Look at the thread 'Two little sailor boys' for an idea of what can happen. |
|
Subject: RE: systematic tune and song structure? From: M.Ted Date: 02 Aug 01 - 04:35 PM I guess I am still not sure what you are looking for here. If you study either verse or music, you'll find that there are sufficient terms for analyzing everything, and at any level. The proof of this is that it is possible for people to learn the rules for writing in any genre, though of course the quality of their work depends on the writer(some of the people around here have managed to write narrative ballads that would pass for traditional, though the subject matter is often contemporary)-- There are a lot of websites that feature the rules of verse, with greater or lesser levels of depth, if that is what you are looking for--as for the musical part of things, the definitive work is most likely Nicholas Slonimsky's "Thesaurus of Scales and Melodic Patterns"-- |
|
Subject: RE: systematic tune and song structure? From: pavane Date: 02 Aug 01 - 04:51 PM Thanks. That will be useful. What I am trying to imagine is a unified notation that I can use to determine by computer program whether a particular lyric will fit a specific tune, or to search for tunes which will match a lyric. |
|
Subject: RE: systematic tune and song structure? From: McGrath of Harlow Date: 02 Aug 01 - 05:20 PM It is possible to sing just about any lyric to just about any tune. You can stretch one syllable over a doizen notes, you can squeeze an indefinite number of extra words into a musical phrase.
Sometimes it sounds brilliant, sometimes it sounds terrible. If you have a computer that knows the difference between brilliant and terrible, you've a pretty smart computer. Even so, different people would differ on a distinction like this, so how's the poor computer to decide who is right. |
|
Subject: RE: systematic tune and song structure? From: M.Ted Date: 02 Aug 01 - 05:20 PM I think there there are a lot of easier ways to do this--the easiest perhaps is just to collect melodies and classify each one based on the structure of the lyric that goes with it, as Burke has explained, this is done in the Sacred Harp Hymnal. Then, when you come across a lyric, you just count the lines and figure out the meter. The thing is that whether you develop a computer program or not, you are still going to have to collect songs and analyze them manually, before you can create a data base. I suppose you could create a program that figures out what the meter and number of lines in a verse is, but you'd have to enter everything syllable by syllable, and notate it in a way that would indicate whether the syllable was accented or un-accented, so you'd still do most of the work off line-- |
|
Subject: RE: systematic tune and song structure? From: pavane Date: 02 Aug 01 - 05:47 PM Yes, I am aware of the data capture problem. And I know it won't be easy - maybe that's why nobody seems to have tried. But the same had to be done with tunes in abc, and look how many of those we have now. |
|
Subject: RE: systematic tune and song structure? From: mousethief Date: 02 Aug 01 - 05:55 PM At a very coarse level, there is the classfication of whether the range of the song runs from tonic to tonic ("authentic") or from dominant to dominant ("plagal"). Alex |
|
Subject: RE: systematic tune and song structure? From: M.Ted Date: 02 Aug 01 - 11:37 PM As I think about it, since ABC notation uses alphanumeric notation that is analogous to melody and rhythmic figures, it might be possible to do something as simple as a "find" on the text files--only problem would be that ABC files are in a variety of different keys, so the same figure would be represented in with different letters, depending on the key--it is an idea, though-- |
|
Subject: RE: systematic tune and song structure? From: JohnInKansas Date: 03 Aug 01 - 01:40 AM Pavane: I believe I may have run into something like you are wanting a few weeks ago, but unfortunately I apparently didn't make very good notes. About all I can recall is that I believe I was looking for Celtic/Irish fiddle tunes, and hit on a fairly large collection. There were some notes indicating "classification" codes for some of the tunes I looked at, and references to "other tunes with same/similar codes; but I didn't dig into the site enough to figure out exactly what he was doing. Depending on how much detail you are looking for, converting everything to abc, and either transposing to a common key or using "semitones up/down/same" would give a good graph of a tune. If I infer correctly what you are looking for, the simple up/down/same notation would not give very good discrimination. You probably need up-how-far/down-how-far/same to be able to get much good out of it. One of the things I was looking at at the time was variant versions of specific tunes. My observation would be that the "frills" would blow almost any system that tried to track all of the notes. There is also the problem of "pickup" notes, that could introduce an artificial "offset" in a simple ID code. Perhaps a coding based just on the "beat" notes would allow sorting down to manageable groups of tunes. I have found the same tune written in Common, CutCommon, 4/4, 2/4, and 6/8 time signatures, and they all sound the same when played. I will try to dig a little deeper in my notes when time permits, and get back to you if I find the site I think I remember. Unfortunately a quick look at recent notes indicates I may not have left myself a trail back. John |
|
Subject: RE: systematic tune and song structure? From: pavane Date: 03 Aug 01 - 02:55 AM Thanks John. I am already aware of the problems of ornamentation. My MIDI file chord analysis program (MIDI Chords) already has 'averaging' code to try and overcome this. It also ignores notes which are very short. You are right in that accenting is very important. There is a new coding system, better than the simple up/down one, mentioned in New Scientist last year, but I haven't managed to get hold of the details. The same tune played in different time signatures is one which I would like to pick up. I think that the Morris Dance Constant Billy is found in both 4/4 and 6/8 in different traditions. The matching which I envisage would match on various different criteria, and perhaps give an overall 'figure of closeness' based on a combination of them.
|
|
Subject: RE: systematic tune and song structure? From: Snuffy Date: 03 Aug 01 - 08:26 AM Bruce Olson has done a lot of work on stress-coding of traditional tunes. His website is in the Mudcat links |
|
Subject: RE: systematic tune and song structure? From: Mr Red Date: 04 Aug 01 - 05:22 AM pavane nothing is impossible but you have a major task here and what percentage susscess are you aiming for? The degrees of freedom on lyrics alone are stupendous. Rhyme patterns, enjambed, internal rhyme, eye/near rhymes, assonance may seem inconsequential but they are the verbal punctuation, the place where we stop & think as audience and breath as performers. You got Rhythm. The grammatical punctuation carries much less weight, the tune dominates as punctuatiuon. Where does the chorus go - is it a chorus or refrain, if it is repeated is it regular? How about choruses that progress, refer back to the verse. AND the famous exceptions. It all adds up to a lot of combinatorial choices (not all of them binary). As an electronic/programming engineer the sheer freedom to songwriting is its attraction for me. As GB Shaw said - re creativity - the Golden Rule is that there are no Golden Rules. Songwriters knew that long before he did. Those who say it can't be done (of anything) are totally correct - within themselves - but I agree this is a mammoth task and destined to be a percentage solution. Good luck |
| Share Thread: |