A Florentine Piano c.1730 for Early Piano Music

This piano with a Cristofori action was constructed in 2003 and is probably the first instrument of this type and compass (56-notes, GG,AA-d³,e³) to have been made since about 1750. It is a fortepiano as Ferrini might have made in 1730 when he was still in Cristofori's workshop and delivered to Queen Maria Barbara in Spain.

Although both original and reproduction pianos of the late 18th century are now familiar to us, the earliest instruments invented by Cristofori around 1700 are less well known since the three surviving instruments are in museums and relatively few reproductions have been made.

The following notes answer some of the frequently asked questions about this fortepiano, explaining its place in the early history of the instrument and some practical details of my work.

A short history of Cristofori, Ferrini, and Scarlatti's involvement with the "Arpicimbalo...di nuova inventione, che fa il piano e il forte".

Scarlatti's connection with the early piano invented by Bartolomeo Cristofori has recently received more emphasis (through the published work of Sheveloff, Badura-Skoda, van der Meer, Tagliavini, Sutherland, Pollens, and most recently from Ogeil) both from the point of view of the music and the instruments. Our knowledge of which pianos and harpsichords were at his disposal in Spain depends almost entirely upon the inventory records made in 1758 after Queen Maria Barbara's death. This list documents four pianos which are specifically mentioned as having been made in Florence, but a fifth piano was probably also of Florentine origin. Thus, a clear link can be established to the designs developed by Bartolomeo Cristofori in Florence from around 1698 (and earlier) to his death in 1732 and Scarlatti's musical activity in Spain. (For further details see David Sutherland, "Domenico Scarlatti and the Florentine Piano", Early Music, May 1995, pp. 243-256).

Cristofori's surviving pianos are dated 1720 (originally FF,GG,AA-c³; Metropolitan Museum of Art, New York), 1722 (C-c³; Collezione degli Strumenti Musicali, Rome), and 1726 (C-c³; Musikinstrumenten-Museum, Leipzig). Since Cristofori was 76 when he died in 1732 we may expect that much of the work in his latter instruments (1720-1732) was carried out by his worker, Giovanni Ferrini. Indeed, Ferrini's surviving instruments show such a strong similarity in the workmanship that this assumption is well supported. A problem in the analysis of all the instruments (i.e. including harpsichords) from the Cristofori/Ferrini workshops has been to distinguish between Cristofori and Ferrini's authorship. Cristofori's first will specified that Ferrini should receive a monetary bequest and the tools from the workshop, so he was the designated successor. Ferrini produced a combination harpsichord-piano in 1746 (Tagliavini Collection, Bologna) which shows that he continued to work with the same action as developed in the 1722 and 1726 Cristofori designs. Thus, for many purposes we can regard the instruments of Ferrini as following in an unbroken tradition from Cristofori's designs and equivalent to Cristofori's own instruments.

The Portugese princess Maria Barbara, Scarlatti's pupil and employer in Spain, may have begun receiving instruction from her teacher during his stay in Portugal from 1719-1723, but he accompanied her to Spain on the occasion of her marriage to the Spanish crown prince in 1729. It is recorded that Cristofori made "instruments" for João V, the King of Portugal, i.e. probably at least one piano and harpsichord, so Scarlatti's prowess at the keyboard may have played a role to the acquisition of a piano or harpsichord. Gerhard Doderer (private communication) has established that Maria Barbara did not bring a piano with her to Spain, so this removal could not account for the four Florentine pianos mentioned in the Spanish 1758 inventory made after her death. In fact there were pianos in each of the royal residences and this suggests that she ordered new instruments when in Spain in order for these to be available, regardless of where she was living.

Carlo Broschi, the famous castrato singer (known as Farinelli), received a bequest from Queen Maria Barbara (on her death in 1758) which according to his own will included a "[cembalo]...à martellino", i.e. a fortepiano. From Burney's testimony, who met Broschi and saw his collection near Bologna in 1770, we learn that the piano was "made at Florence in the year 1730", a year before Cristofori's death. A biography of Broschi by Sacchi appeared in 1784, two years after the singer's death (the text is available at another site through this link). Since this tells us that Broschi owned a fortepiano made by Ferrini it is highly probable that this is the instrument Burney saw and heard Farinelli play. In a manuscript journal (British Museum additional manuscript 35122, edited by Edmund Poole as  "Music, Men and Manners in France and Italy 1770", London 1969,  p. 96, which preceded the published journals of 1771 and 1773 and contains slightly different text) Burney wrote of Farinelli that "...he sings upon it with infinite taste and expression". It is not entirely clear whether Burney meant that Farinelli sang and accompanied himself, or merely that his keyboard playing had the quality of singing. It is practically certain that the fortepiano was the instrument "à martellino" bequeathed to him by the Queen. We may surmise that Farinelli often used this piano, or was accompanied with it, for which reason the Queen left it to him. The date suggests it might have been a wedding present for Maria Barbara, or perhaps the first instrument she ordered when she arrived in Spain. From the available evidence it seems less likely that Farinelli had commissioned the construction of the piano and brought it with him to Spain in 1737.

As Jane Clark has commented (private communication), it is easier to show the connection between the early piano and Farinelli than it is to indicate Scarlatti's involvement. However, in a recent Ph.D thesis, Jacqueline Ogeil has drawn attention to the evidence in music scores of Scarlatti's sonatas, of that which could be best implemented by a piano.

Given that Cristofori died in 1732 and was ill before amending his will in February 1729 it seems highly unlikely that he could have been responsible for building the four Florentine pianos in the years 1729-1730. If Burney's information is accurate then Ferrini probably signed the piano in 1730 before Cristofori died and possibly because he was ill and unable to work. Thus, the likelihood is that at least the 1730 piano was made mostly by Ferrini alone, and that the other three were also his work.

The logic of my reconstruction

Cristofori's earliest surviving piano of 1720 was originally made with an FF,GG,AA-c³ compass (54 notes) and may thus appear suitable as the starting point for a Florentine piano with a larger compass intended for Scarlatti's music. However, it has longer keys than found in the later pianos, which is less advantageous for the player. This increases the inertia of the keyboard and renders the action a little more sluggish since the keylever is the slowest part of the whole action. The 1720 piano also has a style of wrestplank that is inherently weak and is therefore unsuitable for a replica instrument, besides which the present soundboard is not original and dates from the 1938 restoration. In addition the present striking points for the strings are not original and the case appears to have been cut down in length.

In producing a piano for the performance of Scarlatti's music the compass of the two later Cristofori pianos (C-c³) although yielding good evidence of the building practice restricts the choice to some 90 sonatas. With a 56-note compass, as in two of the documented Spanish pianos, one can play some 450 sonatas. Thus, a practical instrument for the performance of Scarlatti must go beyond the strict boundaries of that which we know from the three surviving Cristofori pianos.

I have drawn on the likelihood that Ferrini made many of the pianos at the Spanish court in order to be able to take advantage of the information we can gain from his oeuvre (including harpsichords) and the later developments in Cristofori's piano making. Ferrini's 1746 piano has a compass of GG,AA-e³ (57 notes) and is thus close to the range of two of the keyboards given in the Spanish inventories, which have 56 notes. This might indicate a chromatic compass of GG-d³, but GG,AA-d³,e³ is also possible (i.e. lacking GG# and e flat³). We do not know what the compass was, but Italian makers rarely used a chromatic bass octave and an Italian harpsichord made by Goccini in 1721 (Tagliavini Collection, Bologna) has exactly this GG,AA-d³,e³ compass. A compass starting on GG and reaching to e³ (without e flat³) allows the performance of about nine more Scarlatti sonatas than does a compass ending at d³. The GG,AA and d³,e³ parts of the keyboard also demonstrate a visual mirror-image symmetry.

Since making the instrument I became aware of the piano-making activities of Paolo Morellati in the 1770s who examined Farinelli's piano made by Ferrini. The fact that Morellati's piano had a 56-note range and a compass of GG-d³ suggests that Ferrini's piano may also have had this chromatic range. (See Maria Teresa Nardi, 'I «Cembali a Martellini» di Paolo Morellati', Rivista Italiana di Musicologia, 30 (1995), pp. 373-384.) There is also a harpsichord with the compass GG-d³ in the Stearns Collection, Michigan, which, with David Sutherland's assistance, could be identified from the mouldings as probably having been made by Ferrini (and not Cristofori, as I had previously suggested).

Some details of the instrument

As in Cristofori's 1722 and 1726 pianos there is an una corda stop, activated by sliding the whole keyboard frame by 4 mm to the side. Only one string is struck, the other string of the bichord pair resonates sympathetically giving a slight "halo" to the sound. There is no moderator or any other effect stop.

The action is based on the 1722 and 1726 Cristofori designs that use a hopper mounted in leather guides in the keylever and an intermediate lever which raises a boss on the hammer butt (or rotella as Maffei called it, from his interview with Cristofori in 1709-1710, published 1711). The ends of the cypress hammer shanks are fitted with small blocks on which are glued paper cylinders (a product of 15th-century paper organ pipe technology); these cylinders are the hammer heads. We search for an explanation of Cristofori's motivation for this unusual feature, which is extremely time consuming to make. The paper cylinder is about 1/3 lighter than a solid wood hammerhead and thus yields an action with a slightly lighter touch. There is also a slight audible difference between the sound generated by a paper cylinder or a solid wooden hammer head (both covered with leather), the paper cylinder tending to produce a less brilliant sound when played hard since it deforms and probably thereby increases the area of contact with the string, which increases the damping of the harmonics. The diameters of the cylinders are graded evenly to increase in size from treble to bass, which makes a more noticeable difference to the sound. Larger hammer heads in the bass do not give such a metallic sound as smaller ones.

The configuration of the instrument is with an "inverted" wrestplank (i.e. the nut is on the underside of the wrestplank) so that the hammers impel the string towards the nut thereby giving a more secure contact of the string with the nut. This yields a more stable tone under hard playing and the "limit" of the tonal range is approached gradually rather than there being a sudden onset of the more metallic sound. Furthermore, Cristofori undercut the front edge of wrestplank in the 1726 instrument in order to increase the bearing pressure on the strings on the nut. These design features contribute significantly to the flexibility and range of the tonal resources. Cristofori's attention to these details reveals the nature of the experiments he must have conducted.

Either explanation alone for the use of paper cylinders might be sufficient to reveal Cristofori's motivation, but taking these other factors into account it appears that his aim was to produce a sweet tone, as free as possible of the metallic sound which accompanies hard playing.

Essential to the quiet operation of the Cristofori action is an ingenious "low-tech" solution comprising leather washers in the rotella which damp out any rattle of the hammer butt on the axle rod. The check consists of an independent brass rod and leather pad for each hammer which can be precisely adjusted and is extremely effective. This Cristofori action is obviously the result of much experiment and experience and can be considered as entirely functional.

The 1722 and 1726 actions have relatively short keylevers (shorter than the 1720 design or the later Portugese pianos based on Cristofori's designs) and this allows the speed of the action to be slightly faster. It is faster, for example, than the c.1800 Anton Walter fortepiano (Germanisches Nationalmuseum, MINe 109). When we compare the Cristofori action with Viennese pianos we find the following (the weight to produce a ppp with the dampers in):
Cristofori: 38g (my reproduction at c³; the 1726 action is similar, but less easy to measure)
Viennese: 30.9g (at f³; mean of 8 instruments 1709-1803 reported by Kenneth Mobbs, "A Performer's comparative study of Touchweight, Key-dip, Keyboard Design and Repetition in Early Grand Pianos, c.1770 to 1850", Galpin Society Journal LIV (2001), pp. 16-44, Table 2.)

The soundboard is of cypress wood, a timber apparently not used in pianos after Ferrini's work but typical of Florentine harpsichord making and a material conferring a slightly sweeter sound than the spruce used in other instruments. It also yields a reedier and drier sound in the tenor and bass than one can achieve with spruce soundboard material, partly as a result of the typically higher internal damping of cypress wood. A difficulty conferred by the use of cypress wood with brass strings is that it is more difficult to achieve a singing treble tone and there are usually more "false" notes.

Brass wire is used for the stringing, as the concensus of expert opinion now believes was employed in the original Cristofori and Ferrini instruments, rather than the iron wire found in later pianos. The old brass wire fragments found in the 1746 Ferrini piano/harpsichord are circumstantial evidence, but Morellati also testified to using brass wire in his instrument. This confers what one might describe as a certain "rounded" tone colour in contrast to the normally greater brilliance of iron stringing.

As is now well known, Cristofori developed a double bentside arrangement so that the strain of the strings was taken by one frame and the soundboard attached to a separate frame. In this way case distortion at the bentside did not substantially affect the soundboard, an invention which had to await the introduction of metal frames in pianos (c.1820) before it was even approached in effectiveness. One has to recognise that Cristofori's developed piano was well ahead of the field in many respects and that some of those coming after him produced instruments which were inherently inferior in terms of case stability and the speed of the action.

Some practical considerations in my reconstruction

Although some of Cristofori's pianos have survived, two critical factors determining the quality and volume of sound of the early pianos are not known.

1. The weight of the stringing used by Cristofori is also not known precisely, although we are told by Maffei's article that it was heavier than on harpsichords.
2. A deer leather covering was described by Maffei for the hammers, but the hardness conferred by the tanning process greatly affects the tonal quality and cannot now be established.

Given that we probably cannot determine these unknown factors influencing the tonal performance by documentary research, the only course left to us now is to experiment with the manufacture of new instruments in order to rediscover the possible range of tonal resources at the disposal of the player of the early piano.

From the stringing schemes on Cristofori's harpsichords I have inferred a simple numerical gauge number system he used (see 2000/4 'Principles and Practice...' in my Publications. A version of this Cristofori scheme is implemented in the gauge numbers marked on the 1767 Antunes piano (National Music Museum, Vermillion, SD, USA) suggesting that the Cristofori practice was revealed to Portugese makers through an instrument (or instruments) sent to the Iberian peninsular. This gives us a further, significant clue how to string Cristofori-style pianos. The result of this weight of stringing is that an instrument with a good forte results with adequate power in the treble. (see 2006 'Das Hammerklavier von Bartolomeo Cristofori...' in my Publications). This weight of stringing is also necessary in order to produce a "sweet" tone which does not readily produce a metallic sound on hard playing.

My reconstruction of the case size commenced by seeking to understand the design procedure for Cristofori and Ferrini instruments using their order of thought. It appears that local, Florentine units of measurement explain many of the dimensions chosen, for which reason I define the main case dimensions using them:
Length 90 florentine soldi (= 2479 mm)
Width 34 florentine soldi (= 936 mm)
Height 8 florentine soldi (= 220 mm)

I explain the extreme length of Cristofori's instruments as deriving from the 16th-century Venetian harpsichord-making tradition where the distance between the end of the baseboard at the tail and the wrestplank (in some cases the front edge of the baseboard) along the line of the F string was equivalent to the theoretical length of the F string. In this way the maker established a relationship between the strings (the "soul" of the instrument) and the case (its "body"); see 2008 in my Publications. Cristofori was born in Padova and would thus have learned his craft within the Venetian tradition.

The new width of the instrument derives from the larger compass; this GG,AA-d³,e³ keyboard is nominally 29 soldi instead of the nominally 25 soldi of the C-c³ designs. Cristofori's nominal keyboard width may have been affected slightly by the actual width of the rack holding all the hammer butts. For this reason my hammer rack and keyboard were made before the case, and the case width made to fit it. The case was then built by laying out the stringband on the baseboard, not from a drawing. Specific dimensions were found with the aid of a rule using Florentine soldi. In this way the priority is that parts fit to each other, not that they are made to be exactly x.x mm.

The general tendency is that larger instruments sound rounder and fuller than smaller ones, but the tonal difference between the C-c³ Cristofori design and the GG,AA-d³,e³ size will not be great and the width of the tail is almost the same. As in any instrument with a soundboard, the work done to the dimensions of the soundboard makes the greatest difference. Since wood varies widely in its mechanical characteristics (even in the same plank), the difficulty faced by the maker is always yielding a specific, desired result from the available material.

Although Cristofori developed the double-bentside construction he did not properly understand the principles of structures that are stiffened with bracing. Curiously enough, these principles were understood by timber frame house builders even before the 15th century but the knowledge did not work its way through to instrument making. The modern maker who is informed about structural design and making a reproduction of an old harpsichord or fortepiano is confronted with the choice of repeating an old mistake and inflicting the consequences upon the client, or of avoiding the error. Distorted cases in many large harpsichords and cracks in the treble of the soundboard are examples of such avoidable consequences. I have chosen the route of eliminating the mistakes, which requires very little extra wood in the internal framing, merely locating it correctly. Following this procedure the torsional stiffness of the case (i.e. resistance to twisting) is greatly improved and the soundboard is further isolated from the effects of the strain of the strings on the case.

In the performance of the action there is nothing to improve, however, some materials now available to us are better and our machining facilities are more accurate. For example, I have used a stainless steel rod as the axle for the hammer butts, thus avoiding the corrosion which occurs at wood-brass junctions over longer periods of time.

The result of these procedures and my contribution of experience and skill in preparing the soundboard results in a tone with a full, singing, and "developed treble" (as one player described it) which carries extremely well. In other registers of the instrument a clear tonal character is available. The tuning stability is good so that even after a 1500km transport by car the basic setting of the temperament had not changed.

Review of the Scarlatti concerts in the Musée de la Musique, played by Enrico Baiano and Aline Zylberajch. Diapason, May 2003, p. 60:

Spectacles, vu & entendu
Musique Baroque: Les centenaires et le nouveau-né

La série des concerts Scarlatti de la Cité de la musique a convié la fine fleur des interprètes autour d'un projet audacieux : faire entendre un nouveau fortepiano d'après Cristofori et Ferrini. Reconstruction achevée in extremis (quatre jours avant le premier concert!), marquant l'aboutissement des recherches du facteur Denzil Wraight sur cet instrument mythique - celui de la plus célèbre élève de Scarlatti, Maria Barbara reine d'Espagne -, mécanique à marteaux dans un corps de clavecin italien. On est dès la première note saisi par la rondeur du son, sa projection légère et nuancée, sa personnalité renouvelée sous les doigts des musiciens. Autour du nouveau-né, les clavecins centenaires du musée sont invités à la fête...[followed by reviews of the individual performances]" (Philippe Ramin)