Read Verse Out Loud for Pleasure: The Poetry of Telephone Testing
Keywords: history of audiometry, disability studies, Harvard Sentences, history of quality control, telephone testing
This is an automatically generated transcript of Mara Mills’s paper, ‘Read Verse Out Loud for Pleasure: The Poetry of Telephone Testing’. Although it has been checked, please note that some transcription errors may remain.
Full transcripts appearing online for the duration of Telepoetics (27 May – 5 June 2020) have been reduced to excerpts of approx. 750 words.
[00:00:02] [recording] Her purse was full of useless trash. The colt reared and threw the tall rider. It snowed, rained, and hailed the same morning. Read verse out loud for pleasure.
[00:00:16] This is Mara Mills, Professor of media studies, coming to you from quarantine in New York University, faculty housing in Lower Manhattan. The Harvard sentences have been the gold standard for testing microphones, mobile phones, cochlear implants, synthetic speech and countless other audio technologies since the 1960s.
[00:00:47] [recording] These days a chicken leg is a rare dish. A large size in stockings is hard to sell. Never kill a snake. Never kill a snake. Never kill a snake.
[00:00:51] Currently 72 lists of 10 short, phonetically balanced sentences, they have been picked up in technology journalism and recycled into found poems and automated poetry generators.
[00:01:08] [recording] My name is Max Bartlett and I am from Moscow, Idaho. A king ruled the state in the early days. The harder he tried, the less she got done. The prince ordered his head chopped off. They took the axe and the saw to the forest. The petals fall with the next puff of wind. What joy there is in living.
[00:01:35] [recording] Hi, I’m Marielle Gurnard Willis. I’m from Vienna, Virginia, and this is my Harvard sentence’s poem. Sunday is the best part of the week. A wisp of cloud hung in the blue air. The tree top waved in a graceful way. This is a grad season for hikes on the road. Hoist the load to your left shoulder. Take the winding path to reach the lake. It’s easy to tell the depth of a well. Pop over the fence and plunging.
[00:02:14] The evocative signification of the Harvard sentences obscures their standardized origins and the biases inherent in the normal sampling that produced them. And they are widely miscredited as the start of the test sentence phenomenon, which instead took root in asylums for deaf students, injury compensation tables, industrial laboratories, quality inspections and the mathematical fantasy of average English speech—as I will explain to you in this 30 minute dial-a-lecture. In the 1890s, aurist George Fisk gave a lecture to the Chicago Medical Society titled “The Phonograph in Testing Hearing.” Like many of his contemporaries, Fisk felt that hearing tests were essential for the treatment of ear disease, allowing those in the new medical subspecialty of otology to determine, quote, “one, exactly how much the hearing is impaired as compared with normal hearing and two, whether there is improvement or loss, however slight, between two dates.” New watches of standardized loudness and tuning forks with graduated frequencies had become available to otologists a few decades earlier. But Fiske believed the definition of normal hearing could only be established through a combination of tests with musical tones, watch ticks and spoken words. He wanted what he called everyday hearing to be investigated along with the minimum loudness thresholds for particular pure tone frequencies. But human speakers seemed intractable to the kind of standardization required of a testing tool. Fiske lamented, in his words, “the voice is by far the most important of our tests because it is more necessary in daily life to hear the human voice than any other sounds. But this our most important means of testing, is also by far the most uncertain. Not only is it extremely difficult to reproduce exactly the volume and quality of a given voice, but the voices of different examiners vary so greatly.” Fiske’s solution was to use the new phonograph to turn the voice into a reproducible stimulus. He recorded his own speech on a cylinder and measured the distance at which a patient could just make out the words. He then labelled the cylinder with the date, patient name and hearing distance, and stored it with the patient’s other records to be replayed on the same machine, cranked at the same speed during the patient’s next appointment as a means of assessing the course of hearing loss. Internationally, otolith disagreed that tones were the easiest stimuli to standardise. By the 1920s, relatively quote, “pure tones” with controlled parameters of amplitude and frequency could be generated owing to the development of the wave filter and the vacuum tube in the field of electro acoustics. In 1922, researchers at Western Electric, the manufacturing branch of the Bell system (headed by American telephone and telegraph, AT&T), produced an electronic audiometer that employed a series of tones selected from the piano scale.