Guitar Fingering For Music Performance (Page 4 of 4) in pdf

Recall that both the costs of locality, fret stretch and

any skilled musician. As earlier noted in [5], familiar ﬁn-

vertical stretch and the value of α were determined exper-

gerings may be adopted for new –similar– passages (e.g.,

imentally, on the base of several runs of the implemented

we refer to the practice of the transposition, which on fret-

model, as those leading to the average best results. The

ted instruments is a widespread habit). Chords and arpeg-

gap between the weights obtained is mainly due to the

gios may be part of musical patterns together with melodic

locality factor (that is, respectively 3, 27.75 and 13.5).

fragments, and such patterns may be learned, stored and

Hence, the ﬁrst ﬁngering will be chosen as the predeces-

retrieved as blocks. Cognitive strategies leading to choose

sor for all the ﬁngered positions corresponding to the next

ﬁngerings already known may have suggested the perfor-

note (Figure 3), which is the note D4, and this implies that

mer some ﬁngerings that under the exclusive bio-mecha-

somehow the other two ﬁngerings are pruned. If we don’t

nical aspect are not always preferable, and this could shed

want to loose the ﬁngering solutions starting from (rooted

light on a signiﬁcant part of the incorrect predictions of

in) the second or the third ﬁngering, it is possible to reduce

the model.

the relative strength of the locality by means of a smaller

Future work will address such higher order issues to

α: i.e., by setting α = 0.05, we would yield the locality

further improve the model’s results, and the automatic anal-

factors 0.6, 5.55 and 2.7.

ysis ﬁeld, which can give fruitful contributions to the ﬁn-

gering problem as well.

4. EXPERIMENTAL VALIDATION

6. REFERENCES

We have considered the written ﬁngerings of a guitarist,

[1] Cabral, G., Zanforlin, I., Lima, R., San-

bachelor in guitar performance on three pieces from the

tana, H., Ramalho, G. ”Playing along with

nineteenth Century didactic repertoire of the classical gui-

D’Accord Guitar”, Procs. of the 8

Brazil-

tar: the study n.3, from the guitar method of Aguado (1843),

ian Symposium on Computer Music, Fortaleza,

which is a simple study on chords , the study Opus 60 n.7

CE, 2001.

by Carcassi, which consists of a melody with an arpeg-

gio accompaniment, and the Siciliana Opus 121 n.15 by

[2] Dechter, R. Constraint processing, Morgan

Carulli, which combines chordal and melodic passages.

Kaufmann, San Francisco, CA, 2003.

We have requested the human expert to indicate the

[3] Heijink, H., Meulenbroek, R.G.J. ”On the

phrasing of each piece on the score, and then to mark

Complexity of Classical Guitar Playing: Func-

the ﬁngering that would have adopted in a public perfor-

tional Adaptations to Task Constraints”, J. Mo-

mance. The same pieces were given in input to the im-

tor Behav., 34, 339–351, 2002.

plemented ﬁngering model, with the phrasing annotated.

After the guitarist ended his task, he was requested to in-

[4] Jacobs, J.P. ”Reﬁnements to the ergonomic

dicate whether in any passage the ﬁngering computed by

model for keyboard ﬁngering of Parncutt, Slo-

the model was not practicable.

boda, Clarke, Raekellio, and Desain”, Music

Results. The ﬁngerings computed by the model are

Percept., 18, 505–511, 2001.

recognized to be viable, even though two automatically

[5] Parncutt, R., Sloboda, J.A., Clarke, E.,

ﬁngered phrases were judged very hard to play. This de-

Raekallio, M., Desain, P. ”An ergonomic

monstrates that the set of constraints is adequate at some

model of keyboard ﬁngering for melodic frag-

high level. The correct ﬁngered positions are 90.61% (re-

ments”, Music Percept., 14, 341–382, 1997.

spectively 95.75%, 87.69%, 88.41%). The average num-

ber of missed ﬁngers differences (that is, correct posi-

[6] Radicioni, D., Anselma, L., Lombardo, V. ”A

tion but wrong ﬁnger) has been 9.05% (4.25%, 12.31%,

prototype to compute string instruments ﬁn-

10.59%), and the average number of wrong positions has

gering”, Procs. of the Conference on Inter-

been 2.18% (0%, 4.14% and 2.4%).

disciplinary Musicology, Graz, Austria, 2004,

cim04/.

5. DISCUSSION AND CONCLUSION

[7] Radicioni, D., Lombardo, V. ”Computational

modeling of chord ﬁngering for string instru-

On the whole, the model obtained encouraging results,

ments”, Procs. of the 27

Annual Conference

even though a complete test would require several per-

of the Cognitive Science Society, Stresa, Italy,

formers and a reﬁned metric of distances.

2005.

A closer analysis of the differences reveals that the mo-

[8] Sayegh, S. ”Fingering for String Instrument

del does not still account for some ’idiomatic’ gestures:

with the Optimum Path Paradigm”, Comput.

over the 50% (over the three pieces) of the differences

Music J., 13, 76–84, 1989.

were found on scales and chords played by arpeggio. In

some cases the ﬁngerings provided by the model were

[9] Sloboda, J.A. The Musical Mind. The Cogni-

not strictly worse –under the bio-mechanical aspect–, than

tive Psychology of Music, Oxford, Clarendon

those provided by the expert, but the expert commented

Press, 1985.

that actually they would not have been even considered by

Guitar Fingering For Music Performance Page 4

Related Articles

Related forms

Related Categories