Computerised Violin FIngering Project

[Angela Walker]

Hello, I am a third year computer scientist at Exeter University.

I am also a violin player and have chosen to undertake a project

about violin fingering. The idea is that the violinist

would type in the notes (s)he was trying to play and then an algorithm

would work out the best possible set of fingerings for that particular

passage.

However, I need to do alot of background research and I was wondering if

anyone new of any similar research that was going on anywhere else. If so could you

let me know? Thank you for your time. Angela

[Szilva]

Very interesting project! As far as I know there is no general theory of fingering and no optimal solution therefore it is even more challenging. Possibly interesting links are:

http://n.ethz.ch/student/martense/Software.htm

(I think there is no any algorithm in it) other for guitar, but more interesting:

http://www.nici.kun.nl/Projects/p77/index.html

[some suggestions]

: Hello, I am a third year computer scientist at Exeter University.

: I am also a violin player and have chosen to undertake a project

: about violin fingering. The idea is that the violinist

: would type in the notes (s)he was trying to play and then an algorithm

: would work out the best possible set of fingerings for that particular

: passage.

That's an interesting idea, though I'm not sure if there is such a thing as "the" best fingering for a passage, because on a certain level fingerings tend to be somewhat personal. For example, certain people learn shifting in a certain way, say 1-3-5 1-2-4 (don't worry if you don't know what exactly I meant by that), while others might do it differently, so they would disagree on which fingering is "best" for them.

Nevertheless, I suppose it's at least possible to narrow down fingerings to just 2 to 3 similar alternatives. For starters you might want to investigate some common fingerings for scales and arpeggios of some of the more common keys (eg. C, a, D, Eb, c, E, F, d, G, e, A, Bb, g).

There are so far 2 things I can think of that could make things tricky:

1) slur passages might sometimes require a different fingering than normal in order to avoid awkward shifting in the middle of the slur.

2) certain musical passages might sound best when certain notes are played on a certain string even though it might lead to a more difficult fingering. While sometimes it's explicitly marked in the music that certain notes are to be play on a specific string, sometimes it's not actually written.

There are probably other more subtle issues. So it will be a pretty good accomplishment to at least have it work satisfactorily to simpler types of music, say those from violin pieces and exercises for beginning and intermediate young violinists. Good luck!

[some more suggestions]

Many of the people on this board are busy, but I think it might not be a bad idea to have different people post ideas on certain "rule of thumbs" regarding fingering and see what we get out of the brainstorming. Here's what I've come up w/ so far:

1) for starters, restrict your input to just consider notes (pitch and length) and a key (no slurs, trills, chords, double stops, etc.; nothing fancy)

2) try to fit the scale fingerings for the key you're in to the notes [the "sightreading method" in a sense]

3) if there are certain accidentals (sharps, flats, etc.), try to find the key that best approximates (whatever that might end up meaning) the passage and apply that fingering.

3) avoid consecutive shifts whenever possible (this will especially happen in simplified automating algorithms like the one I've been describing)

4) try to avoid a sustained open D, A, or E.

5) try to keep shifts between long notes or between bow changes. Minimize shifting within fast passages.

Also, instead of thinking in terms of keys, you might also want to consider thinking in terms of intervals.

A huge part of your project will probably end up being trial and error to come up with other rules of thumb. Most likely you'll have a certain basic set of rules, and then on top of it there are a list of exceptions to watch out for.

Once again, good luck!

[Mairead]

The idea is that the violinist would type in the notes (s)he was trying to play and then an algorithm would work out the best possible set of fingerings for that particular passage.

It sounds a fascinating project, Angela. Will you take it further than canonical fingerings, do you think? And do you have a sense yet of how non-deterministic the algo will have to be?

[Irene]

Hi,

I don't know if this will help you or not, since I have never read the book, but it might be interesting.

Author: Flesch, Karl, 1873-1944.

Title: Violin fingering, its theory and practice / Carl Flesch ; English adaptation by Boris Schwarz ; foreword by Yehudi Menuhin. [Hohe Schule des Fingersatzes auf der Geige.] English

Publisher: New York : Da Capo Press, 1979 [c1966]

[David T.]

(1) Within a single phrase, avoid changing string.

(2) A violin is played on scales of just intonation, not tampered tempered scales.

(3) To produce certain tone color, you must play at certain position and/or string.