Designing Effective Speech Interfaces

Computer Hardware

Computer hardware has come a long way since the Harvard Mark I. Howard Aiken and others may not have even dreamed of the level of computing power that personal computers have achieved today. As computer users, we take for granted on a daily basis the complexity of the systems we use. We can also be oblivious to the level of integration we have achieved. Our desktop and laptop boxes are the sum of many parts, all of which are needed to produce computing systems capable of speech input and output. We require sound cards, microphones, speakers, modems, and more to accomplish our speech computing tasks.

Sound Cards

The Altair, released in 1975 by a small company called MITS, was the first personal computer. It had no display and no keyboard. Input was accomplished by setting switches on a panel while output was a series of flashes in a row of lights. As personal computers progressed, the first sounds were made by the Apple II in 1977 using an alphaSyntauri sound card for creating simple crashes and battle noises in fantasy games. Meanwhile, MS-DOS-based PCs could only produce system beeps for a few more years. However, in the 1980s Stanford University developed a synthesis technique called Frequency Modulation, or FM synthesis. Yamaha Corporation purchased and promoted this technology, and the era of digital synthesis began. Previously, analog synthesizers existed, but they were cumbersome, inaccurate, and inefficient.

FM synthesis created new possibilities for employing ...