Harald Bode’s Four-Voice Assignment Keyboard (1937)
Originally published in Contemporary Keyboard magazine (December 1979, p. 89) in Tom Rhea’s “Electronic Perspectives” column. Reprinted in eContact! 13.4 with the kind permission of the author.
Most polyphonic acoustic keyboard instruments share a common characteristic: There is an individual resonating element — string or pipe — for each key. Musical intuition also tells us that the timbral characteristics for a given rank of resonators must be homogeneous, or similar in tone quality. Otherwise, the instrument would have to be restrung / revoiced to anticipate each musical passage having different voice leadings and orchestration. The pipe organ offers some latitude due to its ranks of pies and multiple manuals, but the problem raised by heterogeneous (differing) voices on the same manual are apparent. This type of polyphony must be reserved for an ensemble of performers who play, at any given time, instruments of dedicated tone colour.
Electronics makes it possible to break the first dictum for the keyboard. Now we can produce limited polyphonic instruments having four, five, six, eight voices — that is, a keyboard that has fewer resonating elements (tone generators) than it has keys. But such an instrument must have an assignment system to tune each generator to the appropriate pitch at a given moment. The designer establishes rules for assignment based on geographic considerations or time considerations. A time system gives priority to keys according to the order in which they are depressed. For example, one possible time-priority rule might be stated:
The first generator is assigned to the first key depressed, the second generator to the second key. If the first key is released and then a third key is depressed while the second key is still being held, the first generator is assigned to the third key.
A geographic system has rules concerning the position of notes relative to one another:
The first generator is assigned to the highest key depressed, the second to the next highest, and so forth.
The example is again arbitrary.
On either kind of system, strange things can happen when more keys are depressed than the instrument has tone generators. The extra note may not sound, or it may sound at the expense of “robbing” the tone generator from a previously depressed note. Which note? The designer decides.
One might try to simplify the problem logically: “It’s a four-voice instrument; just don’t play too many notes!” But in the empirical world of experience and observation many discrepancies occur: legato overlap, the impossibility of striking all members of a chord simultaneously, inexact release times and so forth. These problems are further compounded when you have a long release time on the voices. When is a voice through? Does “robbing” cause the note to suddenly disappear when you make a keyboard slip? One might conclude that ten voices would be enough. What about arpeggiated passages such as playing with the sustain pedal down on the piano?
If this is making your head spin, consider the complications when electronics lets us break the second dictum of polyphonic keyboards and have heterogeneous, or different tone colours on the same manual! This idea is particularly tantalizing, for the implication is that the keyboardist will become an orchestra of performers. He won’t. How can a keyboard anticipate voice crossings? How can ten fingers, two feet and one mind provide the many nuances of ensemble playing? The movement of current manufacturers of assignment instruments towards a more complete polyphony and homogeneity of voices is as strong as it is predictable.
Few people are aware that the assignment instrument is over forty years old. Harald Bode struggled with these problems in the mid-30s. In 1937 he produced — in collaboration with a musician named Christian Warnke — the Warbo Formant Organ. Warbo is an amalgam of their names; it was called a formant organ because it used a set of formant filters to produce strong tone colours. 1[1. See the April 1979 issue of Contemporary Keyboard for a discussion of formants.]
A schematic reveals that the Warbo Formant Organ had a geographic assignment system to distribute four voices. The diagram in Figure 1 shows four keys depressed (T1 through T4) and a high-key priority. That is, when a single key was depressed it was assigned to the highest tone generator channel K1. When a second key was depressed, the next highest channel was activated and so forth. The highest four keys would sound if a multitude of keys were depressed — regardless of their order of entry. This meant that new key entries might be disregarded, or might displace other voices: if four keys were held and a fifth key was played lower than those four it would not sound; if four keys were held and a fifth depressed higher, the “new” four-highest notes would play and the lowest key, though still held down, would no longer sound.
Bode was well aware of the musical implications of his assignment system. He installed (on the far left) a column of four switches used to route voices through either of two tuneable filters. The diagram shows voices 1, 2 and 3 routed to the first filter, and voice 4 routed to the second filter. Very different tone colours could be applied — in this case — to a bass / accompaniment configuration.
Bode’s dual filter scheme came from his concept of “complementary tone colors.” In optics, complementary colours have spectra that do not overlap but are mutually exclusive. The analogy in sound is made apparent by the graph in Figure 2: the peak of one filter coincides with the notch (dip) of the other. Two distinctly different tonal colorations result. Of course, when playing through such a fixed filter, one has to avoid pitches that correspond to the notches, since these will have a weak sound. Bode provided separate pedals that were used to control the volume of the two filters — hence the separate voices.
One of Bode’s favourite filter selection schemes was to assign voices 1 and 3 to the first filter and voices 2 and 4 to the second. If you held four keys and then raised only the highest, the three remaining notes would shift up one channel (remember, the highest note is played by channel one, and so forth) . If the filters were tuned to radically different or complementary tone colours, there would be an interesting effect as common tones in chord changes would switch from one channel — and hence one filter — to the other.
The Warbo Formant Organ had no memory device in the keyboard, so notes could last only so long as they were depressed. There were no long release times à la modern synthesizer. But a variety of envelopes were available, and it was possible to play one voice percussively while the other sustained.
It doesn’t take much imagination to see the similarity between the Warbo Formant Organ and today’s limited-polyphony synthesizers. What is ironic is Bode’s spirit of inquiry into experimental voicings and their expressive control as early as 1937. The last card has not fallen on the possibilities for this kind of instrument.