3.3. Instruments selection
At this point, it is essential to decide finally which instruments will be used for the reconstruction. We have already seen that the piano preparations are based on experimentation, and that both the conditions of the piano and the performers taste, make it difficult to standardize their sounds: "the often unclear or even lacking specifications of the mutes in scores make for the variable and undetermined nature of the prepared piano, and aboce all for the problem of subjectivity when describing its acoustical properties" (Vaes, 2009, p. 77). Therefore, based on the fact that it is not possible to relate with precision each piano sound with its corresponding percussion sound , there are some characteristics that can be helpful in making our decisions. Vaes sets a four-degree scale to define prepared pitch sound: "relatively definable pitch" (preparations with rubber, felt, plastic, etc.), "sound complexes with partly indefinable pitch" (preparations with wood, bamboo, etc.), "mostly noise-like" (screw + nut and two screws) and "only noise and no pitch" (bolt + screw). That is, preparations with soft material, such as weather-stripping, fall into the "relatively definable pitch", whereas harder materials, such as bolt or screw with nuts, fall into the "mostly noise-like" group. According to this first approach, Peters edition sounds must have a more defined pitch, and the ones from the original version, less defined and noisier.
By already knowning the general sound characteristic of each version and Cage's collection of instruments, we must find two groups of instruments that produce an homogeneus sound with eleven different pitches (in relation to the eleven preparations of the same material in piano version). One of the groups must have a defined pitch, and the other, a non defined pitch "mostly noise-like". From Cage's instrument collection, we can discard those that produce long sounds (e.g. thunder sheets or tam-tams) because of their lack of clarity in the fast rhythms of Bacchanale. The same would happen with the different rattles. Of the remaining instruments, only drums and cowbells meet the mentioned conditions.
From these options, it seems logical to match the weather-stripping preparations of Peters Editions with drums sound. For the original metallic preparations, the most suitable instruments would be: tin cans, brake drums or cowbells. Although tin cans, brake drums and cowbells can be played with the aforementioned center-edge technique, Cage only used it for tin cans. That is, tin cans is the group of instruments that best fits in the characteristics of the original version.
As we have said before, we will not reconstruct the original version of Bacchanale in this research. At this point we will abandon its approach, shifting our focus to the newest version. As a final contribution before closing this topic, we open the possibility of performing it with tin cans.We can also add another possibility with gamelan instruments, as Virgil Thomson wrote: "the effect [of the prepared piano] in general is slightly reminiscent, on first hearing, of the Balinese gamelang orchestras" (quoted in Vaes, 2009, p. 715).
Back to Perters version, we have already linked the weather-stripping preparations with drums. In 1940, the drums used by Cage were Chinese toms [18], that is, barrel-shape drums with two thick natural skins and no possibility of tuning. Although these instruments have a quite defined pitch and allow the center-edge technique, their tuning can not be modified. Therefore, we could only perform the piece if we had the drums with right tunings. It is extremely difficult to have a set of Chinese drums with all the exact tunings for Bacchanale. For example, in the Royal Conservatory of The Hague there is a small collection of these drums, but only one fits the required tunings. This situation forces us to find an alternative, instruments that replace Cage´s Chinese toms. The best solution would be the use of tamboras [19], also barrel-shape drums with two natural skins. The advantage of tamboras is that they can be tuned with tensioning screws. On the other side, the heads of these drums are slightly thicker than the ones of the Chinese toms, producing a subtly different timbre with less fundamental tone and more overtones.
We have already linked the predominant preparation, weather-stripping with the drums. Now we will focus on the other preparations: small bolt in F pitch and double preparation, weather-stripping + screw with nuts in high B-flat pitch. We will start with the second: in chapter 2.2. "Compilation of alterations and effects" we have explained that, with this double preparation, we listen to both mutes at the same time, and only the screw with nuts preparation when the una corda pedal is depressed. Both preparations sound together only in bars 1, 2, 15 and 16. In our score for percussion quartet, these few occasions are always in second player´s part, without any combination with other sounds in the same part, and written in simple and not too fast rhythms. This allows the player to play two instruments at the same time: one related to weather-stripping preparation and the other, to screw with nuts. According to this decision, B'-flat would be composed of a single sound during most of the piece (indicated in the score with triangular heads). It would be formed by both sound components only in bars 1, 2, 15 and 16. And, in between bar 132 and bar 138, where the una corda pedal is released progressively, the second component would maintain a stable dynamic while the first one would make a crescendo from niente to fortissimo.
It is obvious that the first of these sounds, linked to weather-stripping, must again be a Chinese tom (or tambora). Now we must look for the second component, the one linked to screw with nuts. According to Vaes classification, screw with nuts falls in "mostly noise-like" sounds group. With this preparation, there is still an identifiable pitch, although it is almost hidden behind a metallic sound formed by an attack and between one and three rebounds (depending on the dynamic). A metallic percussion instrument with some small metal fragments on its surface (to work as a small rattle activated by the main instrument´s vibration) would be the ideal choice. We have already discussed the options within Cage´s collection that meet these characteristics: tin cans, brake drums and cowbells. Both brake drums and tin cans do not have a long enough resonance in soft dynamics to activate the rattle. Cowbells are the best solution for this case. We can add to them some small coins (like the pence that Cage´s inserted in the piano), screws, nuts, thin chains, etc. to find the desired sound. The necessary material and its placement must be searched by experimentation once again, because they will depend directly on the kind of cowbell used. We will never use latin cowbells or campanas, much drier and without defined pitch, for this purpose.
Finally, we have to look for the sound that replaces F pitch with small bolt between second and third string. According to Vaes, this preparation falls again in the group of "mostly noise-like" sounds. Its sound is clearly metallic with defined pitch but short, without any rattle, just a slight buzz together with the attack. Perhaps the cowbell option, in this case without the rattle, would be again a good solution. By hitting with the hand, we will get this noisy attack that characterizes the piano preparation.
Once all the sounds are ready, there is only one more question to solve. We know the importance of long notes in Bacchanale, especially when they are at the end of a section, with the question then being: how to coordinate music and choreography? Piano resonance, even with mutes inserted between strings, is longer than drums resonance. For this reason, we have to simulate this effect with a new performative technique, the finger roll used by Cage in his percussion pieces. Cage´s proposal was to roll on a drum with one hand by alternating thumb with the remaining four fingers. However, this kind of roll makes it difficult to hit exactly the same spot, get an even sound and reach the fundamental pitch of the drum. A roll alternating both hands and using one or two fingers of each hand would solve these issues. The notes that can be rolled are indicated in the score with the symbol tr. We must never forget that this techniques have the purpose of simulating the longer piano resonance. That is to say, it is not a matter of keeping a tenuto sound but simulating the progressive decay of piano in the most natural way. This results in small diminuendi during the drums rolls.
The other sound that needs a longer sustain is the F cowbell of the slow section (bars 86, 107 to 119). Taking advantage of the cowbells natural resonance, it is better to look for a beating technique that gets longer resonance instead of using the roll technique. If we use a soft rubber mallet (similar timbre to our hands) but quite heavy and we hit very close to the edge, we will get a longer sound and a slower decay.
In case of very slow tempi performance, in which we need more sustain in the slow section, we can use felt mallets on the drums and F cowbell (excluding the spots where we use rubber mallet). In order to keep the noisy sound in B-flat cowbell, we will continue beating it with the hand.
To conclude the reconstruction process, it is necessary to verify that all the elements and techniques were used in the first pieces for percussion of John Cage. Our set-up has Chinese toms (or substitutes) and cowbells, both instruments widely used by the composer. About the techniques: hands hitting center and edge, fingers rolls and rubber mallet, Cage also used all these techniques. In Annexe IV. "Recordings" we can find recordings of each part where the instruments, tunings and techniques are showed.
3.2. Percussion score
Due to this diagram, the original piano score can be divided into four players. The result, being the attached score for piano quartet in which the original pitches and notes are kept as in the original version (Annexe II. "Piano quartet score"). However, we cannot play this score with percussion instruments. At this point, we should choose the instruments we will use to perform it in order to write a new score for them. Instead of choosing a fixed list of instruments, we will work with some general facts and leave the final decision until later in the process. We know that Cage did not use pitched percussion instruments in his early works, so our palette of instruments will be restricted to unpitched instruments. If we are more specific about it, the instruments Cage used were, most of the time, drums (with natural skins), traditional instruments from other cultures (rattles, teponaxtle, etc.), metal instruments (thundersheets, cowbells, sleigh bells, etc.) and common objects (tin cans, brake drums, sheets of paper, etc.). The purchase of instruments such as "brass gongs, brass cymbals, chinese toms or wood-blocks" (Shultis, 2009, p. 96) for his own collection is justify by invoices in the name of Cage.
In chapter 2.2. "Compilation of alterations and effects", we mentioned the different preparations between the first performance of Bacchanale (1940) with metal objects inside the piano, and the preparations required in Peters Edition score edited thirty years later (mostly felt preparations). There are at least, two ways of preparing the piano for Bacchanale, resulting in two completely different sound palettes. We could transfer this idea to our reconstruction, meaning to work on two versions of the piece: one closer to the original preparations and the other one to the newest preparations. However, due to the lack of knowledge about the preparations used in 1940, as well as the lack of recordings, we cannot have a precise idea about how this first version sounded. Because of this situation the difficulty to choose the most suitable instruments for its reconstruction becames bigger. For these reasons, we will propose options for its performance and interpretation, but they will not be put into practice. The goal will be to perform only the Peters edition.
In any case, for both piano versions of Bacchanale, the performer uses the same score and the same instrument interface (piano keyboard). In our reconstruction, we will keep this fact: the score will not change, neither the instruments placement, although they will produce other sounds. Therefore, a percussion ensemble can play both versions by just changing the instruments, but they will not need to learn a new score or disposition of instruments.
Then we need to create a percussion score. We already have the division for four players and instruments set-up, but we still have to make it understandable for percussionists. In images [14-17], we can see how many instruments each percussionist plays. Some of the instruments are used with the two sounds technique (center-edge), but not all of them. To make the score reading as understandable as possible, each line will represent one instrument. However, we will use a round note head to designate the middle sound and a cross note head for the edge sound. Instruments with only one sound are written with round heads. A triangular head designates instruments which do not belong to the drums group. The following images [14-17] shows the notation for each percussionist. Annexe III. "Percussion quartet score" shows a transcription for unpitched percussion instruments.
3.1. Part division and instrument setting
Once we have enough knowledge about Cage´s early music and his piece for solo prepared piano, we can start to reconstruct it. The first question that we have to answer is: how many percussionists will be required to perform it? Most of Cage's percussion music composed between the 1930´s and the 1940´s is written for percussion ensemble (3-6 players), and, as we said before, the piece was planned originally to be played by Cage´s percussion ensemble. Additionally, the big size of percussion instruments and the chords with more than four notes make it impossible to play with only one percussion player. Therefore, a percussion ensemble playing Bacchanale sounds like the most logical idea. This is just a partial answer to our question: we know that more than one percussionist will be needed, but we still have to find out the exact number of performers.
For the moment, we assume that each percussionist can just play two sounds at the same time, one with each hand (i.e. no four mallets technique or similar). In this case, we need at least three performers to play the five-sound chords (bar 1). However, if we consider the instruments and set-up sizes and the amount of sounds to be played by each percussionist, we will realize that a version with four performers is more suitable. We, therefore, need to divide the original score between four players.
The best criteria to distribute the sounds is to keep, when possible, the melodic lines played by one musician. That means various performers will need the same sounds in different parts of the piece. For example: bar 1, at least two players are required to play the chord pitches G, C and F; however, two bars later these three notes are part of the same melodic line. At this point, we have two options: to duplicate sounds (each player will have all of the sounds in his own set-up) or to share sounds (put all the instruments together in a unique set-up and performers will play it from different sides). The advantage of duplicating sounds is that each performer can set the instruments at his own taste, but on the other side, it is extremely complicated to get exactly the same sound (timbre and pitch) in two percussion instruments. Due to this, Bacchanale would work better if we keep each original sound related to its corresponding percussion instrument. Therefore, our set-up will have as many percussion instruments as different pitches/preparations has the original score: twelve.
However, if we pay attention to the pitches used by Cage to compose this piece, we will see that some of them can be grouped together in a chromatic scale: A, B-flat, B, C, D-flat, D, E-flat & G, A-flat. In Cage´s percussion music, it is common to play two different sounds from each drum, tin can or any other object. One sound is produced playing in the middle of instrument or skin and the other, playing on the edge. This second sound is slightly higher than the first one. Since we are reconstructing the piece, it seems logical to use this Cage way of composing and performing percussion music in our version. That means we can assemble each chromatic interval in one instrument and its two sounds. Following these criteria, our final percussion set-up will have this appearance (players 1 & 3 are in front of players 2 & 4) [13].
