Abstract: This article traces the history of speech wave visualization and the longstanding relationship between phonetics, communication engineering, and deaf oral education. American telephone engineers drew on this history to build the sound spectrograph in the 1940s, a machine that transformed the representation of sounds by considering speech not in terms of meaning nor in terms of airborne waveforms but in terms of the characteristics of its perception and the minimum features by which it could be reconstructed. The sound spectrograph was designed to make telephone transmission more efficient and to support deaf oral communication; the ability of deaf subjects to read spectrograms was, moreover, the best evidence for the identification of information-bearing features in a complex speech wave. The sound spectrograph directly influenced information theory, which gave mathematical instructions for the efficient digital encoding of audio and visual signals. Spectrograms suggested that much of the content of speech was redundant or irrelevant and could be discarded without a listener perceiving any difference. It will be argued that deafness ultimately served as an "assistive pretext" for nineteenth-century phoneticians and twentieth-century engineers, who quickly turned to more profitable applications for their devices.