Also found in: Dictionary, Thesaurus.
sound spectrograph[′sau̇nd ′spek·trə‚graf]
a device for obtaining and recording images that show the variation in time of a spectrum of complex sounds, including the sounds of speech. The main idea of the sound spectrograph is the representation of the speech process in the form of a plane graph, in rectangular time-frequency coordinates. The intensity of each component of a sound of given frequency at a given moment of time is reflected by the density of blackening of a sensitive layer of film or photographic or electrochemical paper or by the light intensity of a phosphor.
The recording type of sound spectrograph is used for the spectral analysis of transient (variable) sounds. These devices consist of a recording and reproducing section and an analyzing section. The first part of the device is a drum tape recorder, which makes possible the recording of the sound to be analyzed during a certain segment of time (several seconds) and the subsequent unlimited reproduction of the sound. The analyzer of the device consists of an adjustable band-pass filter with a continuously variable central frequency, which permits the sequential study of all of the frequency components of a given sound. The recording device produces blackening of sensitive paper that is stretched over the drum; the blackening is proportional to the filter output voltage, which in turn increases with the strength with which the frequencies corresponding to the given pass band of the filter are expressed in the signal. A “line,” which changes in density according to the variation in time of the spectrographic intensity of the sound, is recorded on the paper during each revolution of the drum, which is mounted on the same shaft as the drum tape recorder. By reproducing the sound being analyzed a great number of times, each time changing the pass band of the filter, a series of successive lines is obtained that displays a pattern of the variation of the spectrum in time.
Sound spectrographs in which the image is received on a moving phosphor layer are used in experimental linguistics and in teaching practice for the study of foreign languages, the training of deaf-mutes, and the correction of speech defects. In this case, a sound that is picked up by a microphone is continuously analyzed (that is, it is dissected into frequency ranges) by means of a multichannel frequency response analyzer, each of whose channels has a band-pass filter, amplifier, and miniature quick-response gas-discharge lamp. Each lamp illuminates its own track on the continuously moving tape, which is coated with phosphor. Upon movement of the tape, the spectral intensity in each frequency band is represented by the brightness of illumination of the phosphor on the corresponding track—that is, the image in this case is also obtained in the form of lines. Because of the afterglow of the phosphor the tracks remain visible during the period when the image is moving across the “window” of the device (approximately 1 second). The de-vice enables a specially trained individual to directly “read” speech that is spoken into a microphone at a rate of 90-120 words per minute.