To run the TypistInstrument, click the applet link above. You need to install JSyn to run this:


The Origin and Outline of the Idea


This applet is my first musical instrument for typists. The keys that are pressed trigger sounds, the words and sentences that are typed build the structure of the composition. A score could be recordings of various interviews or even live performers speaking, having their words transcribed. The live performers could be reading a poem, for example, while an ensemble of typists transcribed the words and sounds into music to accompany the poem.


The idea for this project began with a discussion with a friend of mine who had a job as a typist. She sat in a large room full of other typists, all of them transcribing recordings of police interviews. Its very detailed work, there are many rules about how to transcribe things like pauses and sounds that are not words (things like "um", or "oh"). I remarked that this process is almost like someone listening to music and transcribing to musical notation. She inserted that these typists have a well established skill, much like a musician. And through this discussion, we came up with the idea of turning these typists into an orchestra.


Since keyboards output distinct key events with distinct values and durations, much like a piano, it seemed obvious that the Typist Instrument should be a designed to make tonal music. In the first rendition, I directly mapped the key code for each key to a pitch. This lead to a very dissonant clang of sound most of the time. In next version, the ability to set the musical key was introduced. It is set by the first key typed, with the letters A thru G directly setting that musical key. For any other keyboard key, the musical key is set by taking the modulus of 12 of the keyboard key code. The vowels were mapped to the tonic (a,u), mediant (i,y), and dominant (e,o) of that scale. For the rest of the keys, the key code was directly used to generate the pitch. This was a minor improvement, but still mostly sounded like a continuous clang.

So for the final version, a new mapping scheme was introduced. The eight notes of the scale were mapped to the most frequently occurring letters in written English: ETAOINSR. They were mapped roughly in order of consonance with the tonic, with R mapped to the octave of the tonic. For the next most frequent letters, HLDCUMF, the modulus of 12 of the key code was taken, giving notes in the same octave as the previous keyboard keys. For the rest of the keyboard keys, the key code was directly mapped to pitch number.

There are three instruments being played. The first is a stylized typing sound which is played upon each key press. Each word typed generates a chord that is played through the second instrument, an enveloped square wave. All of the pitches from the letters of the word are sounded at the same time to generate the chord. And whenever punctuation (,.;:!?) is typed, a phrase is played thru a filtered sawtooth. That phrase is an exact copy of the phrase as what was typed, but transposed down two octaves and given a random amplitude.


Ideally, a piece of text would be composed in conjunction with the TypistInstrument so that when that text is typed, composed music is performed. This text could then be read aloud while the typist transcribes the spoken words, thereby performing the composition. The TypistInstrument could also be tailored to a specific text so that a recognizable text would produce new music. And lastly, a number of different TypistInstruments could be designed to create the orchestra of typists.

(C) 2004 Hans-Christoph Steiner, released under the GNU GPL