Speech synthesis confers a number of benefits to technology end users. It allows individuals with impaired eyesight to be able to operate radios and computers. For those who cannot speak, and who may also have trouble using sign language, speech units such as the device employed by Stephen Hawking allow a person to communicate in ways unthinkable a century ago. For these individuals speech synthesizers play an integral role in adding quality to their day to day lives. On our local repeaters synth voices make announcements about nets and club events, and speech synths read the weather on the NWS frequencies. Beyond these specialized uses, one of the ways everyone can share in the joy of chip talk is through the medium of music.
The IBM 704 was the first computer to sing. It was first introduced in 1954 and 140 units had sold by 1960. The programming languages LISP and FORTRAN were first written for this large machine that used vacuum tube logic circuitry. Bell Telephone Laboratories (BTL) physicist John Larry Kelly coaxed the 704 into singing Daisy Bell aka A Bicycle Built for Two using a vocoder program he wrote for the 704.
Lovely as the a cappella computer was, it was deemed in need of instrumental accompaniment. For this part of the song the expertise of fellow BTL employee Max Vernon Mathews was sought out. Max was an electrical engineer whose first love of music enabled him to become a pioneer in electronic and computer music. In 1954 he wrote the first computer program for sound generation, MUSIC, also used on the IBM 704. The accompaniment to the voice portion of Daisy Bell was programmed by Max in 1961 using the IBM 7090.
The IBM 7090 was the transistorized version of the 709 vacuum tube mainframe. The 7090 series was designed for "large-scale scientific and technological applications." The first of 7090's was installed in late 1959 at a price tag of close to $3 million. Adjusted for inflation the price today would be a whopping $23 million buckaroos. Besides its musical capabilities, the 7090's other accomplishments included being used for the control of the Gemini and Mercury space flights. IBM 7090's were also used by the Air Force for the Ballistic Missile Early Warning System up until the 1980s. Daniel Shanks and John Wrench used it to calculate the first 100,000 digits of pi. Yet none of the above uses compare, in my mind, to the beauty of the IBM 704 joining forces with the IBM 7090 on the song Daisy Bell.
Another computer, HAL 9000, still gets most of the credit for this electronic version of Daisy Bell. Arthur C. Clarke, author of 2001: A Space Odyssey, happened to be visiting his friend and colleague John Peirce at BTL when John Larry Kelly was making his demonstrations of speech synthesis with the IBM 704. He was so fascinated by witnessing this computational marvel that six years later he wrote that version of Daisy Bell into his screenplay, as sung by HAL in the middle the machines climactic mental breakdown. The song was on the vinyl platter "Music from Mathematics" put out by the Decca label a handful of decades ago (listen to video above.)
Daisy Bell went on to have a notable reprise for the Commodore 64 when Christopher C. Capon wrote his program "Sing Song Serenade". The sounds for his version were played direct on the hardware by rapidly moving the read/write head of the computer. . The resulting audio was emitted from the floppy disk drive.
Max Mathews continued to make strong contributions to the humanities in the realms of music and technology. In 1968 he developed Graphic 1, a graphical system that used a light pen for drawing figures that could be converted into sound. In 1970 Mathews developed GROOVE (Generated Real-time Output Operations on Voltage-controlled Equipment) with F. R. Moore. GROOVE was the first fully developed music synthesis system for interactive composition and realtime performance. It used 3C/Honeywell DDP-24 (or DDP-224) minicomputers.
An algorithm written by Mathews was used by Roger M. Shepard to synthesize Shepard Tones. These tones (named after Roger) consist of a superposition of sine waves separated by octaves. When the base pitch of the tone is played moving upward or downward, it is known as the Shepard Scale. Playing this scale creates an auditory illusion of a tone that continually ascends or descends in pitch, yet seems to get no higher or lower. It is the musical version of a barber pole or of the Penrose stair, a type of impossible object in geometry, made famous in the drawing Ascending and Descending by M.C. Escher.
Max also made a controller, called a Radio-Baton and Radiodrum, used to conduct and play electronic music. Developed at BTL in the 1980s it was originally a kind of three-dimensional mouse. The device has no inherent sound of its own, but produces control signals that are used to trigger sounds, sound-production, effects and the like. The Radio-Baton is similar to a theremin. Magnetic capacitance is used to locate the position of the conductors baton, or mallets in the case of the drum. The two mallets are antennas transmitting on slightly different frequencies. The drum surface, also electronic, acts as another set of antennas. The combination of these antenna signals is used to derive X, Y and Z, and these are interpreted according to the assigned musical parameters.
Many of these mainframe musical programs are now available in the program Max that can run off a laptop.
Besides the use of Daisy Bell in the soundtrack for 2001, director Stanley Kubrick used a wide range of work by modern composers. The piece Atmospheres written by Gyorgy Ligeti in 1961 was used for the scenes of the monolith and those of deep space. Ligeti's earlier electronic work Artikulation, though not used in the film, shares an interesting connection to some of the ideas behind speech synthesis. Artikulation was composed in 1958 at the Studio for Electronic Music of West Deutsche Radio in Cologne with the help of Cornelius Cardew, an assistant of Karlheinz Stockhausen (whose works involving shortwave radios will be explored in time). The piece was composed to be an imaginary conversation of multiple ongoing monologues, dialogues, many voices in arguments and chatter. In it Ligeti created a kind of artificial polyglot language full of strange whispers, enunciations and utterance.
Music from Mathematics: Played by IBM 7090 Computer to Digital Sound Transducer, Decca LP 9103.
Gyorgy Ligeti: Continuum / Zehn Stucke fur Blaserquintett / Artikulation / Glissandi / Etude fur Orgel / Volumina, Wergo 60161, 1988.
Justin Patrick Moore
Husband. Father/Grandfather. Writer. Green wizard. Ham radio operator (KE8COY). Electronic musician. Library cataloger.