| 摘要 |
Apparatus which provides a real time sound spectrogram of signals such as speech which are complex in their frequency content and also vary rapidly in time, is described. A visual display device such as a scan converter and video monitor is provided which receives, in short time segments, signals from a time compression heterodyne type spectrum analyzer and displays these segments on a raster. Display control circuits generate a vertical sweep signal and fast and slow horizontal sweep signals in synchronism with the short time segments. Successive spectrums are plotted and displayed under the control of the fast horizontal sweep and the vertical sweep while the spectrogram, as of an entire speech utterance, is made up of successive spectrums during the interval of the slow horizontal sweep. Thus a sound spectrogram of each short time interval is produced; there being only a short delay after utterance of the sound before the result appears on the display. The intensity of each frequency in the spectrum during each time segment and over each spectrum (viz., each piece of the spectrogram) is plotted in real time as the speech utterance is produced. The pieces are concatenated such that the perceptual effect is that of a continuous sound spectrogram which unfolds in real time. The frequency-time-intensity plot thus contains fine detail enabling the viewer to observe the characteristics of the speech utterance of interest such that the apparatus is especially suitable for use in speech training as to deaf students. |
