As a musician
and a noder
, it strikes me as somehow inadequate that E2 as of yet lacks any efficient way to transcribe music into text
. After all, how can any database that claims to be everything
possibly be complete without commercial jingos
, popular song melodies
, famous bits of classical music
Some noders have tried producing ASCII staves, representing musical notes and flags as so many |'s, /\'s, and O's, to mixed effect. Generally, producing such representations is so labor-intensive (not to mention, potentially confusing, since there is really no way to distinguish between half notes and quarter notes without lapsing into special characters), it doesn't make any sense to use that as the primary means of transcription of music.
A possible solution occurred to me the other day as I was reviewing my old music anthology the other day; the first piece was an ancient Greek piece, called the "Epitaph of Seikilos," which was really a drinking song by a fellow named Seikilos, inscribed on a tombstone.
The interesting thing about this particular song is that it features musical notation that doesn't require a staff. Pitches are represented by letters, and rhythm is notated using special symbols, inscribed over the notes they are connected to. Thus, a full song (even with the words under their corresponding notes) is transcribed into text and symbols, and can be translated back from the notation on the tombstone to a musical staff, and re-played.
Heck, if the Ancient Greeks could do it, why can't the musicians of E2?
In the interest of beginning a discussion leading to a possible solution to the problem of a music-less E2, I will begin by setting out some beginning principles that I feel an efficient system of musical transcription must follow.
First, the system must be simple
By this, I mean, using as few characters as possible. Using special characters increases the likelihood that another noder will only see a bunch of broken symbols. So we stick to what's available on your standard QWERTY keyboard.
Further, I think it would be best if we used a minimum of characters. Using too many characters, say, using "8" to mean eighth note, using "ET" to mean 8:3, using "(" to mean slur this note, etc., would mean that any proficient transcription reader would have to learn a lot of arbitrary symbols just to read music. This would be a signficant barrier to the use and re-use of our transciption method.
Second, the system should be founded on principles
One of the simple beauties of modern musical notation is that it is built for the most part on principles that transcend spoken languages. For example, a pitch is indicated by a black dot, placed somewhere on a staff of five lines, relative to the placement of a clef and accidentals. Its duration is indicated by the absence/number of flags attached to the note. Once you understand these principles, almost all of the music in this notation is readable.
Contrast the ease of reading notes to translating the words that often accompany them. As any classical musician will know, it won't do just to read the notes; there are usually Italian words included to indicate how one should play a given passage. And learning Italian isn't sufficient, either. Play Debussy or Ravel, and you'll have to know French. Play Wagner, Strauss, or Mahler, and you'll have to know German. You can also find music with English and Russian, depending on the publisher. And never mind if the composer chooses to use performance indications that are non-standard.
I wish I was multilingual. Oh well.
So, in the same spirit, our musical notation should be based upon principles, as well, which can be expanded on without a massive amount of explanation or rote memorization. Using this E2 musical notation should be nearly intuitive once the basics are learned.
Okay, so here's a start:
First of all, I think it would make sense to represent notes using their standard letter names:
C C#/Db D D#/Eb E F F#/Gb G G#/Ab A A#/Bb B etc.
Another character will be needed to indicate rests. I think "$" would be good for that. I don't know why.
Taking from the Greeks, we could represent music in two lines of information, the bottom indicating pitch, the top rhythm.
Meter and Key Signatures
As to the question of meter and key: I think, for meter, it would be best to remember our theoretical names for different meters. For example, 3/4 would be "simple triple" meter, 6/8 would be "compound duple," etc. Recall that "simple triple" actually applies to a lot of different meters: 3/2, 3/4, 3/8, etc., as do the other combinations of (compound, simple) with (duple, triple, quadruple). But since we don't have flags or stems to worry about in this forum, there is no longer any need for us to indicate what note receives the beat in our notation of meter. That a beat occurs is given. How many beats occur is what still needs to be notated, and what that beat's primary subdivision is should be notated.
By primary subdivision here, I mean whether the meter is compound or duple. Which is to say, in compound meters, the beat's primary subdivision (you can also think of it as the "default" subdivision) is into three parts, and in duple, it is into two.
So, conceivably, any music which is in a compound duple meter can be indicated by "2/3" to mean, "two beats in a measure, primary subdivision into thirds." That would cover 6/4, 6/8, 6/16, etc. So the first number indicates the number of beats in a measure, the second, the primary subdivision.
Key signatures could be indicated in short hand; for example, a key signature containing F#, C#, and G# can be indicated by "3#," meaning three sharps. Similarly for flats. Since the order of sharps and flats is always the same, owing to the circle of fifths, no more information is necessary besides the number of accidentals. A "natural" key signature could be indicated simply by "0." We also need an ASCII "natural symbol." How about "n?"
The remaining difficulty is the notation of rhythm, which can be complex and detailed. Well, if we are again to look to the Greeks, they indicated the "beat" with no notation at all. Two beats received one kind of symbol; primary subdivisions, another. So let's say:
- Notes that receive a beat's duration have no symbol
- Notes that receive the primary subdivision receive a "/"
- If the note is the first subdivision of the beat, but is a subdivision different than the primary subdivision (e.g. eighth note triplets in a duple meter), it receives a "\"
- Notes that receive the secondary subdivision receive a "/" or "\", depending on whether the subdivision is duple or compound
- Further subdivisions can be notated using a combination of "/" and "\", depending upon the subdivision of subdivisions
- "Dots" -- indicating that a note's duration is increased by one half -- receive a ".", effective on all preceding slashes for that note
- Ties are indicated by "-"
- Notes of duration greater than a single beat receive a "_", followed by a "-" and further notation depending on length:
- Another "_" for every additional beat
- A "." to indicate a "dot", effective on all preceding notation for that note
- Standard notation for notes tied to the first note (i.e. "/" for a primary subdivision, "//" for a secondary notation, etc.)
- Any ties (indicated by "-" as needed)
- Bars are separated with "|"
- When non-standard subdivisions of the beat occur, each note receiving such a subdivision receives a "\" (indicating non-primary subdivision), along with a superscript containing a ratio indicating the number of such subdivisions to the number of beats those subdivisions would cover. When more than one of these occurs sequentially, they can be enclosed in "( )" and receive the superscript outside the right parenthesis. For example, triplet quarter notes in a simple duple or quadruple meter could be indicated: (\ \ \)3:2 Quintuplets: (\ \ \ \ \)5:1
- Grace notes -- or really, all "small notes" can be indicated with a "`"
- Trills might be best indicated with a "!" after the note name, instead of on the line for rhythm
I doubt that is sufficiently exhaustive or simple. But this is just a starting point.
Standard note-names will become ineffective if the music we're transcribing goes beyond an octave in range, since they won't indicate which octave some notes are in. One solution would be to adopt the same names that some theorists have given to all the notes, according to their position on the grand staff, with numbers, various capitalizations, apostrophes, etc.
Originally, I thought it would be best to indicate such octaves by superscripts and subscripts, but as it turns out, such characters don't monospace well, so an alternative arrangement could be to use the following rule:
A primary octave is designated by the transcriber. The lowest note of this primary octave is indicated in parentheses after the key signature (e.g.: 4b(G)). All of the notes in the octave above that note will be designated by plain, regular-formatted upper-case letter names.
Beginning with the next occurrence of the note that is the base of the primary octave, we use nested formatting to indicate octaves, in the order of italics, bold, underline (e.g. The G above the aforementioned G would be G, the next G above that G, the one above that G).
The octaves below the primary octave can be designated by lower-case letter names. The first octave below the primary octave would be simply lower case; then we use nested formatting for notes below that, again, italics, bold, underline (e.g. So the first G below the primary octave would be g, the next, g, g, g).
For clarity's sake, it would be best to limit formatting to the note name itself, and not to its accompanying accidental, to avoid possible confusion in the case of lower octaves of Bb.
This provides us with eight octaves of notation, which should be sufficient for any melody, anyway. If music transcription gets to the point where we need more (that is, if we ever get to the point of transcribing full pieces of music, instead of melodic fragments), this will probably have to be amended. It's an imperfect rule, I think, but it maintains the ability to monospace, which I think should be preserved as much as possible.
Oh, and MONOSPACE
This should make transcription over two lines easiest, so that you don't have to account for variable character length when trying to match slashes to note names.
That's about all I can think of, so far. A few examples, to demonstrate:
Twinkle, Twinkle, Little Star:
2/2 | | |_-_| | | |_-_| | | |_-_| | | |_-_|
0(C) CC|GG|AA|G |FF|EE|DD|C |GG|FF|EE|D |GG|FF|EE|D |
2/2 | | |_-_| | | |_-_||
0 CC|GG|AA|G |FF|EE|DD|C ||
Oboe Solo from Tchaikovsky's Fourth Symphony, 2nd Movement
:(I would have chosen a classical example more well-known, if I had any suitable scores handy)
2/2 // // |/ / //|// / /|/ / //|/// / |/ // /|/ / // |
2/2 / // / |/ / // |/ / //|// / /|/ / //|/// / |// //|
2/2 // / /|/ / / / |/ / // |/ / / / |/ // / |// / /|/ / ||
Something a bit more complicated, perhaps? Fine!
Oboe Solo from Beethoven's Third Symphony, 2nd Movement:
4/2 _-_-_/.//|_-_`/.///.//|_-_ |_-_///.///.|_-_ /.//|
3b(G) $ G G |G DC BnC D |Eb C$|G F EbD C |Eb DD D |
4/2 _-_/.///.//|_-_-_-_-_ | /.// ||
3b D D EbF G |Ab |AbFEbD|EbEbEbEb$||
Well, that's a start. I'm curious to see what changes other musicians on E2 might have. There are several weaknesses with this notation, weaknesses I haven't totally solved. But I am relatively confident that this can be readable, with enough practice.