Denzil Wraight - Italian Keyboard Instruments |
PITCH IN ITALIAN STRING KEYBOARD INSTRUMENTS
Discussion of the pitch of Italian harpsichords, which began in the 1960s with the work of Shortridge, Barnes, Hubbard, and Russell was complicated from the beginning by a lack of agreement between the researchers on some basic matters of stringing and instrument design. In addition, the limited data on original scales and compasses at that time prevented a better appreciation of the problems. Pitch is important not only for the musicological issue of describing the place of the harpsichord in musical life, but also because the type of string material used (brass or iron) has a significant tonal effect.
Italian string lengths covered a wide range from about 150 to 420 mm, measured at c², but the main range for 8' instruments was between about 255 mm and 360 mm. The short end of the whole range comprises mostly small virginals and some clavichords. At the longer end of the range we find exceptionally low-pitched harpsichords.
The situation in Italy
A striking feature of many Italian compasses is that they end on either c³ or f³. Shortridge (1960) linked this observation with the scalings and suggested that the compasses C/E-c³ usually had short scales (averaging 266 mm) and that the C/E-f³ compasses were associated with long scales (averaging 327 mm). This data led to the interpretation (which we can call the 'transposing hypothesis') that harpsichords and virginals with C/E-f³ compasses were pitched a fourth lower than those having a C/E-c³ compass, by analogy with the Ruckers 'transposing' harpsichord which had keyboards at pitches a 4th apart. Similar arguments were advanced by Barnes (1965) to support the same idea that Italian instruments were made at two different pitches in order to facilitate transposition.
Thomas and Rhodes (1967) argued that the difference between the groups of string lengths was simply due to the use of iron wire for long scales and brass wire for short scales. A contrary view, that all Italian instruments were strung with brass wire with the pitches proportional to string lengths was later advanced by Barnes (1968; in Ripin, 1971, and again in 1973) and defended against the views of Thomas and Rhodes. Other contributions by van der Meer (1968), Tagliavini (1974), Hellwig (1976) and O'Brien (1990, and unpublished work) considered various aspects of these problems.
It has been claimed that documentary evidence demonstrates how Italian harpsichords and virginals should be strung (Hellwig, 1976), but closer study shows that all documents tend to be vague on some important aspect. Documents alone cannot answer the questions, although they can change the balance of the arguments when used in conjunction with evidence from instruments.
After a comprehensive study of instruments and documents I brought forward much new evidence and published conclusions on which the following text is based (see my dissertation cited below which summarises earlier contributions in detail and includes an extensive bibliography). The discovery of original scales (despite alterations) and the attribution of unsigned work has made a substantial difference to the amount of data from which conclusions can be drawn.
The significance of string lengths
It might appear as if string lengths were a matter of personal choice until it is appreciated how regularly and accurately instrument makers used the same scales in their instruments, and different workshops used the same scales. The twelve instruments (harpsichords and virginals) made by Celestini in Venice are a good example of the regularity of the scales used by an individual maker, but other workshops in Venice agreed in using closely-defined string lengths.
A possible latitude of pitch for any given string length was reported in earlier sources (Russell, 1959, p. 32; Hubbard, 1965, p. 9; Shortridge, p. 103, and MGG, 1995, p. 497). This view arose because it is possible to tune a string over a range of pitch below its breaking point and still produce a result which would satisfy some listener. However, some of the best known Venetian harpsichord makers also built organs, such as Lorenzo da Pavia (for further information on Lorenzo as an organ maker at this website click here), Alesandro and Vito Trasuntino, and Domenico da Pesaro. The physical obligations and constraints of pipe length for the organ builder were thus known in the same workshop where harpsichords were designed. It remains yet to show the exact extent to which the organ pipe length in Venice influenced the choice of string length for a given pitch. Nevertheless, it seems that there was a fairly fixed relationship between pipe length and string length, string length and pitch (see Wraight, [in press] for further discussion).
Some later modifications to 16th-century instruments also show that makers regarded it as desirable to alter the scale of an instrument even when a pitch change of only a semitone was involved. One instrument by Vito Trasuntino (dated 1560, Staatliches Institut für Musikforschung, Berlin) was rebuilt with a shorter scale. This gives the impression that the pitch was raised, but when the different string materials are taken into account, the later (brass) scale lowers the pitch by almost a minor whole tone (ratio 10:9, 182 cents). Since the instrument was originally made for the higher end (a' = 465Hz) of the 'normal' 8' pitch range (a' = 465Hz to 415Hz), the modification brought it nearer the lower end of this range. The unmodified pitch of the instrument would probably have been called mezzo punto and the modified pitch tuono corista (see below).
There is strong 18th-century evidence from instruments and documents for the use of brass and iron wire. The traditional relationship between scale and pitch is also clearly shown in some instruments by Cristofori, those he influenced (Ferrini and Solfanelli), and others, which indicate a ratio of close to 5:6 for the lengths of brass wire and iron wire at the same pitch (see O'Brien, 1981 and Wraight 2002/1). However, the range of scales found before 1600 would allow for the use of brass and iron wire. The problem is to identify the appropriate stringing material for each instrument.
Assigning pitches to scales: virginals
Of all the instruments under consideration, it is easiest to assign iron stringing to the small virginals with c² scales of about 150-170 mm. These would then have been octave virginals, as they are also described in documents, to iron scales of about c² = 300-350 mm which would have sounded from a' = c.415Hz. to c.465Hz.
More complicated is the case of 8' virginals where documentary evidence is vague and sometimes contradictory. One polygonal virginal ('Artus Gheerdinck 1603', National Museum of American History, Washington DC) now attributed to Antegnati c.1540 used two sections for the right hand bridge so that the ratio of scales is close to 6:5. This appears to reveal the intention of using separate bridges for iron strings (in the treble) and brass strings (in the bass). Other virginals with c² scales of about 300-350 mm would also have been strung in iron wire so that this whole group of instruments was tuned at a 'normal' 8' pitch (i.e. a' = c.415Hz to c.465Hz).
Those virginals with a c² scale of about 270-290 mm present a different problem: in recent years it has become widely accepted that brass strings are appropriate for short scales. However, my analysis of the design of the ratio of treble to bass strings in virginals of all sizes concludes that the short-scaled virginals were intended for iron strings at a higher 8' pitch (a' = about 520 Hz; Wraight, diss., Part 1, pp. 209-212). This apparently unusual pitch has a parallel elsewhere since O'Brien has already argued for the iron stringing of the 1548 Ioes Karest virginal built in Antwerp, an instrument which is similar to those considered here (O'Brien 1990, pp. 23-26).
There are few 16th-century virginals which were designed for brass strings; the intarsia of Isabella d'Este's virginal (ducal palace, Mantua) being one of them, for which documentary evidence has survived that supports the possibility of brass stringing.
Assigning pitches to scales: clavichords
16th-century Italian clavichords were also designed for iron strings, mostly at the higher fourth or the octave pitch. Only one instrument (Musikinstrumenten-Museum, Leipzig, no. 3) has been designed on different scale principles and may have been intended for brass strings (Wraight, diss., Part 1, pp. 215-224).
Assigning pitches to scales: 16th-century harpsichords
Most of the 16th-century harpsichord scales were originally relatively long (c² = 300-350 mm) and many of these instruments had a 4' stop and a compass C/E-f³. The available evidence indicates the use of iron strings. The documentary evidence of Galilei (1581, p.143; first noticed by Ripin, in his unfinished work for the entry 'Harpsichord', in The New Grove Dictionary of Music, 6th ed.) indicates that the 'gravicembalo' was strung with iron strings in the treble and brass wire in the bass, although how far into the bass the iron stringing extended is not specified. The scale design of these harpsichords requires brass wire only for the last few notes and implies that iron-strung, long-scaled harpsichords would have stood at normal 8' pitch (i.e. a' = about 415 Hz; Mitchell ascribes a much lower pitch to these scales of a' = about 348 Hz, which is controversial. See also Wraight, 2002/2).
Scales remained surprisingly constant in Venice throughout the 16th century, but our knowledge of the strength of wire at this period is scanty. Thus, our appreciation of the intended pitch for these early instruments may gain in precision in the coming years.
Assigning pitches to scales: lower-pitched harpsichords
Further complications are introduced by the fact that documents refer to harpsichords 'a fourth low' (alla quarta bassa) and that the name 'gravicembalo' might appear to suggest a low-pitched instrument. Later usage of the name 'gravicembalo' does not indicate a low-pitched instrument, although it might originally (c.1500) have indicated a harpsichord at low pitch, i.e. 8' pitch with respect to the prevailing 4' pitch of chamber keyboard instruments.
There are some harpsichords with even longer scales where c² = 410-470 mm (e.g. 1574 Baffo, Victoria and Albert Museum, London; 1561 Franciscus Patavinus, Deutsches Museum, Munich; 1579 Baffo, Musée de la Musique, Paris) which would have been pitched a fourth lower than the c² = 300-350 mm scales even if strung with iron wire.
A few harpsichords have a long scale (the c² is mostly about 300 mm) even when they neither have a 4' stop nor reach to the f³ note. The Italian tradition of scale design indicates that these were also intended for iron wire strings, at the same pitch as the virginals having the same scales, i.e. pitched a tone above the scales where c² = 350 mm. Indeed, some harpsichords with c² = about 300 mm (albeit with a 4' stop) are simply proportionally-scaled versions of lower-pitched instruments (e.g. '1585 Bortolotti', Musée Instrumental, Conservatoire Royale, Brussels, at the lower pitch and c.1563-1570 Dominicus Pisaurensis, Stiftelsen Musikkulturens Främjande, Stockholm, at the higher pitch). This reflects the preference in 16th-century Venetian instrument making for the use of iron strings (probably dating back to the 15th century), regardless of instrument type, size, or compass.
Assigning pitches to scales: higher-pitched harpsichords
Harpsichords with short scales (c² = 270-290 mm) could appear (when only the c² scale is considered) to be intended for brass wire at normal 8' pitch. My analysis of the scale design (including the bass strings) implies that probably all of these harpsichords were intended for the high pitch in iron wire. Thus, they would stand at the same high pitch found for some of the 8' virginals (a' = about 520 Hz). Some examples are the 1521 Hieronymus Bononiensis (Victoria and Albert Museum, London) and the 1584(?) 'Rigunni' (Stearns Collection, Ann Arbor, Michigan). The 1554 Dominicus Pisaurensis harpsichord (Musée de la Musique, Paris) has longer bass strings than many of these instruments and appears much like the 17th-century design of harpsichord, but considered alongside the other Dominicus harpsichords it is an iron-wire scale design. This is inferred because it is a proportionally-scaled version of the 1543 Dominicus octave-harpsichord (Musée de la Musique, Paris) which was intended for iron wire strings.
The 1554 Dominicus harpsichord shows that without special knowledge there may be no clear criteria for distinguishing between the scaling designs for brass and iron-strung harpsichords of this size. Thus, we cannot be sure of the intended pitch of some 16th and early 17th-century harpsichords, and this problem may in practice be irresolvable.
On the probable assumption of iron-wire stringing, a few harpsichords were pitched a fourth above normal 8' pitch, (1608 Celestini, Kunst und Gewerbe Museum, Hamburg; 1596 Celestini, Royal Ontario Museum, Toronto). Two harpsichords by Dominicus Pisaurensis (1543 and 1546, mentioned above) were made at 4' pitch, the highest pitch in normal use in the 16th century.
Thus, we may summarise for Venetian instruments as follows: all long scales (c² = 300-350 mm) can be seen as prima facie indications of iron wire strings, regardless of the type of instrument or compass. Short scales (c² = 260-300 mm) in 16th-century virginals are usually also evidence of iron strings. However, short scales in harpsichords before 1600 may have been intended for iron wire in some examples and brass wire in others. The string material used was not exclusively linked to a type of instrument, type of compass, or even a type of scale (i.e. long or short).
The 17th century
Throughout the 17th century short scales from all cities (c² = 250-300 mm) predominate and in most cases these indicate 'normal' 8' pitches intended for brass wire stringing (roughly a' = 465 Hz to 390 Hz). However, some may have been intended for the high 8' pitch (i.e. a' = c.520 Hz, strung with iron wire), but it is difficult to distinguish this design from the brass-strung harpsichord. Indeed, it is possible that what we have come to appreciate as the 'traditional' Italian harpsichord with a short, brass-wire scale is in fact merely the re-stringing of a 16th-century, iron-wire scaled design.
The 1610 Vincentius harpsichord (Vassar College, Poughkeepsie, N.Y.) might have been intended for the high 8' pitch (i.e. a' = about 520 Hz), strung in iron wire, but the 1637 Zenti bentside spinet is a clearer example that this high pitch was still used well into the 17th century. Another exception is the 1628 Albana harpsichord (c² = 248 mm, Museo Civico, Bolgona) which was probably intended to stand a fourth higher than 'normal' 8' pitch using iron strings. The long scales (c² = about 350 mm) at 'normal' 8' pitch intended for iron wire, which were common in 16th-century Venice, are unknown in harpsichords of this period, although used in virginals until the 1630s. A few harpsichords, even as late as c.1680 - c.1700 (e.g. c.1680 Migliai, Germanisches Nationalmuseum, Nuremberg, MIR1078 with a faked inscription 'Domenico da Pesaro') have c² = c.300 mm, which, when strung with iron wire, corresponds to the higher of the two 8' pitches normally used before 1600, i.e. a' = about 465 Hz. Thus, the scales used, and therefore the pitch levels, after 1600 were substantially the same as those of the 16th century and no obvious rise in pitch took place.
There is little evidence to explain why brass wire scales at the same pitch as iron wire scales should have been preferred in the 17th century, but it may have been the result of a preference for a bolder and louder sound. That brass wire gives a louder, even coarser sound was well known at least as late as 1511 since Virdung records this (1511, fol. Fi), and the clear tendency towards a 2 x 8' disposition in the 17th century also suggests the desire for volume of sound. However, it must also be remembered that this analysis of a change of string material is based on the comparison of mostly Venetian scales before 1600 with largely Florentine and Roman scales after 1600. It is possible, but not yet proven (and may be unprovable), that there was always a strong tradition of using brass scales in Rome and Florence.
In discussing pitch it is necessary to deal with the specific situation in each centre of music making, rather than speak of 'Italian instruments'. This is required, if for no other reason, simply by the amount of information available which differs widely from city to city. Very few Roman virginals survive and evidence of harpsichord making is mostly from after 1620. The predominance of short scales (c² = 263-280 mm) for the harpsichords suggests brass stringing at a 'normal' 8' pitch of a' = about 415 Hz. Some longer scales used by Albana and Ridolfi (c.300 mm) could indicate a slightly lower pitch if strung with brass wire and this would accord with documentary evidence Doni gives, according to which pitch in Rome was lower than further north on the Italian peninsular (discussed by Haynes, pp. 73-81).
There are some documentary references to harpsichords with an 'ottava bassa': Urbani and Zenti made such instruments, and since both worked in Rome this may have been a Roman speciality. One of Zenti's harpsichords has survived, albeit having subsequently been given a faked 'Cristofari' inscription and reduced in length. It appears to be identical with the first instrument mentioned in the Medici inventory of 1700 and was made at 16' pitch (see Wraight, 1991, Gai, 1969, pp. 6-7).
Neapolitan & Sicilian instruments
The longer scales we find in most of Guarracino's virginals (c² = 290-300 mm) could be taken to imply a pitch at the lower range of 'normal' 8' pitch, i.e. a' = c.400 Hz to c.390 Hz; this assumes a stringing in brass wire. It would bring the pitch levels into general accordance of what Doni tells us, so that Neapolitan pitch was no higher than in Rome. The pitch of Grimaldi's two surviving harpsichords could be assigned to the a' = 415 Hz to 400 Hz range. When other Sicilian instruments are identified our view will become clearer.
The evidence of 18th-century harpsichord pitches is mainly from Florentine harpsichords, and it is probable that the whole tone range of 'normal' 8' pitch was in simultaneous use (e.g. the c.1680 Migliai strung in iron wire at about a' = 465 Hz), although scales of c² = 250 mm are first found from 1731 onwards (see Sutherland, 1998-1999, Table 2, p. 16). Thus, there is no clear indication from these instruments that pitch rose in the 18th century compared with earlier times, rather that the range of pitches for which instruments were made remained at a constant level from the 16th-18th centuries.
Implications of the pitches
As mentioned above, evidence of 16th-century scales is mostly from Venice: instruments were designed for two main 8' pitches a tone apart, as well as quart and octave instruments. The high 8' pitch (a' = about 520 Hz) is also clearly indicated by a number of harpsichords and virginals. This scheme of pitches is similar in structure to that described by O'Brien for the Ruckers instruments built in Antwerp; this is unlikely to be accidental and reflects the same use of instruments.
It can be seen that the scales of Venetian instruments reveal a degree of coordination in the manufacture which previously had hardly been suspected, although this need not imply that the use of the instruments was as well organised. There were enough intermediate sizes of instruments that pitch incompatibilities when performing could easily have arisen.
Most of the virginals and harpsichords with C/E-f³ compasses sounded at 'normal' 8' pitch (a' = 415 Hz to 465 Hz), or higher. Only a few harpsichords were actually at a low pitch and the 'transposing hypothesis' that all instruments with C/E-f³ compasses were at low pitches can be discarded. The high f³ sounded a high pitch and would have facilitated the performance of music at octave pitch (printed music rarely went into this range). Most small 15th-century chamber keyboard instruments were at octave pitch. That other compasses were C/E-c³ does not necessarily indicate any special pitch relationship vis-a-vis C/E-f³ compasses, but simply reflects the Italian tradition of extending the musical range only in steps of a fourth (or fifth) and much less frequently by individual notes (e.g. from c³ to d³). It is partly this step in the end notes of compasses from c³ to f³ which has been responsible for confusion because it appeared as if it might be connected with the pitch of the instruments.
Following Haynes' examination of the terms tuono corista, tutto punto, and mezzo punto (Haynes, pp. 56-72), it seems possible to identify them with widely-used string keyboard pitches in Venice and the north of Italy. I have therefore referred to these on this website and give the following approximate pitch equivalents (Haynes' values are given in brackets):
The value of 415 Hz which I have given here is to be understood as approximate; the pitch might in practice have been down to about 400Hz, or such a pitch might have been regarded as equivalent.
BibliographyBarnes, John, 'Pitch variations in Italian keyboard instruments', GSJ 18 (1965), 110-116.
Barnes, John, 'Italian string scales', GSJ 21 (1968), 179-183.
Barnes, John, 'The specious uniformity of Italian harpsichords', Keyboard Instruments: studies in keyboard organology 1500-1800, ed. Ripin, E.M. (Edinburgh, 1971), 1-10.
Barnes, John, 'The stringing of Italian harpsichords', Der Klangliche Aspekt beim Restaurieren von Saitenklavieren', ed. Schwarz, V. (Graz, 1973), 35-39.
Gai, Vinicio, Gli strumenti musicali della corte medicea e il museo del conservatorio 'Luigi Cherubini' di Firenze (Florence, 1969).
Grove: see New Grove.
Haynes, Bruce, Pitch standards in the Baroque and classical periods, Ph.D. thesis (University of Montréal, 1995).
Hellwig, Friedemann, 'Strings and stringing: contemporary documents', GSJ 29 (1976), 91-104.
Hubbard, Frank, Three Centuries of Harpsichord Making (Cambridge, MA, 1965).
O'Brien, Grant G.,'Some principles of eighteenth century harpsichord stringing and their application', The Organ Yearbook 12 (1981), 160-176.
O'Brien, Grant G., Ruckers: a harpsichord and virginal building tradition (Cambridge, UK, 1990).
Meer, John H. van der, 'Harpsichord making and metallurgy-a rejoinder', GSJ 21 (1968), 175-178.
MGG: Musik in Geschichte und Gegenwart, Allgemeine Enzyklopädie der Musik begründet von Friedrich Blume, ed. Finscher, Ludwig (Kassel, 1995). 'Cembalo' by John H. van der Meer.
Mitchell, Nicholas, 'The 1531 Trasuntino Harpsichord In A Universal European Pitch System', Harpsichord & Fortepiano 9 (2001), 7-13.
New Grove Musical Instrument Dictionary, ed. Sadie, Stanley (London, 1980; 2/2001).
Russell, Raymond, The Harpsichord and Clavichord (London, 1959; 2/1973).
Shortridge, John, 'Harpsichord-Building in the 16th and 17th Centuries', United States National Museum Bulletin 225 paper 15 (1960, 2/1970), 93-107.
Sutherland, David, 'The Florentine School of Cembalo-Making Centered in the Works of Bartolomeo Cristofori', Early Keyboard Journal 16-17 (1998-1999), 7-75.
Tagliavini, Luigi F., 'Considerazione sulle vicende storiche del corista', L'Organo 12 (1974), 119-132.
Thomas, William R. and Rhodes, John J. K., 'The String Scales of Italian Keyboard Instruments', GSJ 20 (1967), 48-62.
Wraight, Denzil, 'A Zenti harpsichord rediscovered', Early Music 19 (1991), 99-101.
Wraight, Denzil, 'The stringing of Italian keyboard instruments c.1500 - c.1650', Ph.D. dissertation, Queen's University of Belfast, 1997 (UMI order no. 9735109).
Wraight, Denzil, 2001/2, 'The pitch relationships of Venetian string keyboard instruments', Fiori Musicologi, Studi in onore di Luigi Ferdinando Tagliavini nella ricorrenza del suo LXX compleanno, edd. Seydoux, F., Castellani, G., and Leuthold, A. (Bologna, Editore Pàtron, 2001), 573-604.
Wraight, Denzil, 'Some Evidence for the Use of Brass and Iron Stringing in Italian Keyboard Instruments, Early Keyboard Journal 20 (2002), 91-106.
Wraight, Denzil, 2002/2, letter to the editor, GSJ 55 (2002), 403-404.
Wraight, Denzil, [in press] 'Establishing Pitch Relationships in Italian String Keyboard Instruments', Proceedings of the conference "Stimmton und Transposition im 16.-18 Jahrhundert', Bremen October 1999.