The earliest speech synthesizer was created by a Russian professor, Christian Kratzenstein (1723-1795). Between 1773 and 1779 Kratzenstein made acoustic resonators and produced vowel sounds by connecting them to organ pipes.
A contemporary in Vienna, Wolfgang von Kempelen, produced a more advanced machine in 1791. His "acoustic mechanical speech machine" was able to produce single sounds and even words or short phrases. He is best known for an earlier invention, a chess-playing machine named "The Turk." This consisted of a cabinet, housing (apparently) just cogs and wheels, and a manikin with movable arms. One could not see the legless human chess player concealed inside. Once this hoax was exposed, his legitimate speech machine was discredited as well.
Alexander Graham Bell became interested in speech synthesis after he saw a replica of one of Von Kempelea's speech machines. When young, Bell had taught his pet terrier to stand between his legs and growl while he manipulated the dog's vocal tract by hand. He was eventually able to produce "How are you, Grandmamma?"
Joseph Farber improved on Von Kempelen's machine by adding a mechanical tongue and a pharyngeal cavity that could be manipulated as well. It was powered by bellows and controlled by a keyboard and could sing as well as produce speech.
The first electrical speech synthesizer was the VODER, developed by Homer Dudley and presented at the 1939 World's Fair. It saw more utility as the VOCODER, which reduced ordinary speech into a facsimile to reduce the bandwidth necessary for the telephonic transmission. This allowed a larger number of telephone calls to be transmitted over a line.
