Sound Judgment

One of the most incredible bits of by play of the Middle East war was the Israeli-intercepted radio-telephone conversation between Egypt's President Nasser and Jordan's King Hussein, as they conspired to save face by blaming their disastrous defeat on U.S. and British air intervention. But was the identity of the voices firmly established? To test this point, London's Daily Telegraph submitted a recorded tape of the Nasser-Hussein talk to U.S. Physicist Lawrence Kersta, president of Voice-print Laboratories, Inc., in Somerville, N.J. Along with the tape went a two-year-old CBS News recording of what was known to be Nasser's voice.

"100% Sure." After running the Israeli tape through an electronic filtering process to eliminate static, Kersta chose 25 basic phonetic elements from the voice believed to be Nasser's, and the same elements from the CBS recording. These "phonemes," as they are called, included such sounds as ah, ee, eye, o and yeh, which are common to both English and Arabic. Words from the tapes containing the phonemes were then fed into a spectrograph, which electronically translated them into signals that activated a stylus.

Moving rapidly over a strip of paper on a slowly revolving drum, the stylus traced out distinctive patterns, or voiceprints, that were determined by the frequencies, loudness and duration of each of the phonemes. Finally, after a night in which he painstakingly compared the patterns produced by phonemes from the two tapes, Kersta concluded that they had all been uttered by the same person. He reported to the Telegraph that he was "100% sure" that the voice on the Israeli tape was that of President Nasser.

Convicting Evidence. Kersta's conclusion--and his voiceprint technique--is based on the principle that every individual's voice is as unique as his fingerprints. Because the frequencies and energy distribution of the human voice are determined by the size and coupling of the nasal, throat and oral cavities and by the manner in which each person uses his articulators (tongue, teeth, lips, soft palate and jaw muscles), Kersta says, it is highly improbable that any two voices can be identical. Thus, voiceprints, like fingerprints, can be used to make a positive identification. Whispering, muffling the voice, changing its pitch, or even mimicking another voice will not alter the basic voiceprint pattern.

Since he developed the voiceprint system at Bell Telephone Laboratories in 1962, Kersta has worked with law enforcement agencies on more than 100 cases involving voice identification. His voiceprints were used to convict a rioter in the Watts area of Los Angeles who, with his back to the camera, admitted to a TV interviewer that he had set fire to five different buildings. Last year voiceprints were admitted as evidence in a jury trial in New York.

Voiceprints have also received implicit recognition by the State Department, which sent a Middle East expert to help Kersta examine the Israeli tape. But Washington had good reason to believe that the tape was authentic even before Kersta's analysis: neither Nasser nor Hussein ever denied that the recorded voices were theirs.

This file is automatically generated by a robot program, so reader's discretion is required.