Monday, November 30, 2009

IBM 704 Speech Demostration at Bell Laboratories (1962) recorded by D.H. Van Lenten


"HELLO LADIES AND GENTLEMEN" those words were spoken by an electronic computer. They are an example of synthetic speech, a product of Bell Telephone Laboratories Research into the basic nature of speech. Knowledge developed through such research may be useful in devising new techniques for transmitting speech over communication systems. To make the computer talk, it is fed punched cards containing the names of speech sounds, the computer combines these sounds in accordance with the linguistic rules which govern the English language, into connected intelligible speech. For example when the sounds for the sentence "He saw the cat" are fed into the computer in sequence it says "HESAWTHECAT", the flat monotonous tones of the computer indicate an absence of the pitch and timing characteristics natural to human speech. When timing information is added to the punched cards, the computer says "HE SAW THE CAT". The sentence still sounds unnatural but when pitch information also is added to the cards the computer speaks and accents almost completely human except for a slight electronic twang. In the following sentence most of the variables inherent in human speech have been specified on the punched cards, the computer makes one of the pivotal remarks in the development of the telephone. "MR. WATSON COME HERE, I WANT YOU" The present quality of speech synthesis by computer is illustrated in the next example, part of a famous soliloquy from Hamlet. "TO BE OR NOT TO BE, THAT IS THE QUESTION; WHETHER 'TIS NOBLER IN THE MIND TO SUFFER THE SLINGS AND ARROWS OF OUTRAGEOUS FORTUNE, OR TO TAKE ARMS AGAINST A SEA OF TROUBLES, AND BY OPPOSING END THEM? TO DIE, TO SLEEP" Singing in purely physical terms is essentially a matter of pitch and timing, in the next selection the computer sings a familiar ditty;
DAISY, DAISY, GIVE ME YOUR ANSWER, DO,
I'M HALF CRAZY ALL FOR THE LOVE OF YOU.
IT WON'T BE A STYLISH MARRIAGE
I CAN'T AFFORD A CARRIAGE,
BUT YOU'D LOOK SWEET UPON THE SEAT
OF A BICYCLE BUILT FOR TWO
The computer now sings the same tune, but to a musical accompaniment played by another computer. Piano students will notice that the music producing computer has a rather stylised left hand. Incidentally synthesizing music on a computer is almost as formidable as making a computer talk. To get the samples of synthesized speech we've heard so far a computer's memory was stored with 34 speech sounds and a set of rules for producing these sounds and for making the transitions from one sound to another. When the computer was fed the names of speech sounds on punched cards it was in effect told what to say. But its manner of saying it, even its dialect and apparent accent was determined by the rules stored in its memory. The objective of this program is to formulate a minimum set of rules for making plausible english speech. The next two selections however were produced by analyzing a person's speech and reconstructing it synthetically on a computer. The objective here is to duplicate the sounds and transitions made by a human speaker, including his dialect and accent. With such a program the computer sounds like this; "MEN STRIVE BUT SELDOM GET RICH" "NOW IF YOU ALL DON'T MIND, I'D LIKE TO SAY A FEW WORDS ABOUT TEXAS" And now the computer would like to express its appreciation for your attention. "THANKS FOR LISTENING"

Vocals programmed by John Larry Kelly Jr. and Carol Lockbaum
Musical accompaniment programmed by Max Mathews

No comments:

Post a Comment