The introduction of the software sampler in the last ten years has enabled sound designers to create electronic copies of historic keyboard instruments, utilizing the enormous memory capacity and data processing speeds of the personal computer. When the proper acoustic instrument is selected, prepared correctly, and sampled with state-of-the art techniques, and when the sample editing is well performed and the proper playback system is used, the sounds are impressive, especially if the player is skilled. What is more, a variety of voices are availableall voices of the harpsichord or organ, for example, in a typical electronic keyboard instrument. The instrument stays in tune and costs much less than its acoustical equivalent.
It has been argued, by both acoustic and electronic musicians, that the sound of an electronically sampled instrument is inferior. If so, the simplest explanation is that the sound emanates from one or more loudspeakers, rather than from strings or pipes. In the absence of a side-by side comparison, however, the difference may be of little consequence to many listeners. Understanding such distinctions as trade-offs, there are advantages and disadvantages to be counted on each side. There may be interest, therefore, among both players and listeners, in the development of an electronic system offering a greater diversity of keyboard sounds than is currently available.
The goal of the present proposal is to make readily available to any user the voices of an historically rich collection of keyboard instruments, including several not currently available, such as the 18th century clavichord and fortepiano, as well as the various Steinway models of the 19th century. Specifically, the vision is to create a free-standing keyboard instrument, housing a dedicated computer, in a system enabling the player to change instruments as easily as changing stops on an organ.
Who, one may ask (besides the exceedingly small number of early music enthusiasts who are also interested in electronic reproduction), would comprise the public for such an instrument? A little imagination produces several categories of potential users: music educators and teachers in the classroom; music historians, anthropologists, or musicologists interested in comparing the sounds heard by audiences of other eras to those of our own; composers and music experimenters; performers using historic instruments in unusual situations (for example, outdoor concerts without rain shelter); and students (or parents) who cannot afford acoustical keyboards because of their high price or considerable size.
The large companies with the means to develop such a musical instrument appear unwilling at present to do so, most likely because they are not convinced that sales would ever be great enough to make a profit. The project itself, however, is highly feasible, for instance at a university, given the presence of an interested, interdisciplinary team of researchers, including both faculty and students.
The first task is straightforward: identify the historic instruments to be sampled, find examples of each and get the cooperation of the owner and of a recording engineer to do the sampling. Electronic sampling is a combination of art and technology, and Michiel Post in Amsterdam, Netherlands, is an acknowledged expert. His disk, "Post Historic Keyboards," already contains a virginal, two two-manual harpsichords and an 18th-century Anton Walter fortepiano. To complete the 18th century, a clavichord should be sampled. There may also be interest in including a "reconstructed" fortepiano as well as the existing "restored" fortepiano in the collection.
The 19th century may be covered with three or four pianos, including the Erard piano and the Steinway piano of 1860, credited as the first modern piano. None of these 19th century instruments has yet been sampled. Michiel Post would be the logical choice to do this, as the instruments are all in Europe (there are few if any distinctive 19th century pianos in North America). The important 20th century pianos include the John Cage Prepared Piano and many other modern pianos, all of which have already been sampled. In summary, the number of instruments to be sampled is five or six, and the total number of sampled instruments available to the project would be 10-12.
The second part of the project is the conversion of the samples to CD-ROM. This is a highly technical process, but many successful examples exist, including several produced by Michiel Post. The "product" at this point would be a virtual library of the sounds of 10-12 sampled keyboard instruments spanning the 18th, 19th, and 20th centuries.
As a first step, the conversion should be done in a standard format that can be read by an existing synthesizer or computer. If this path leads to an acceptable final product, it may not be necessary to invest heavily in the third major part of the project, namely the design and construction of a special-purpose computer. What is needed in terms of computer capacity is the ability to link physical keyboard "input" (that is, by the player) to the sampled instruments on the CD-ROM, and to allow the player to switch among the samples via buttons or switches on the front panel. The computer hardware would accept a "MIDI in" connector from the physical keyboard and have a "line out" port for the audio system. To allow for additional samples, the front panel should have perhaps 15 buttons. Ultimately, the entire set up - the keyboard, an easily manipulated and robust interface for switching samples, the computer, and the audio system - would be housed together to comprise a single playable instrument.
A university research team, including faculty and students, of electrical and computer engineering and also faculty and students of music then would study the instrument for its musical and practical value and begin the process of refining and adapting the system as needed. One could imagine the team including a faculty member or graduate student who wished to research the change in sound of keyboard instruments over time, or wanted to demonstrate how a particular piece of music sounds when played on the instrument for which it was composed. In such cases actually playing an instrument, instead of using recordings, offers a certain immediacy both to the musician and the audience. As the system is refined for any such application and becomes easier to use, additional copies might be made, or new variants developed.