VISIBLE PATTERNS OF SOUND—POTTER 205 
training and rehabilitation. This has proven true in the case of the 
congenitally deaf engineer noted above. During the past year his 
speech has improved considerably and can be understood quite well by 
the average person with whom he comes in contact. His work with 
the translator patterns, however, has emphasized the desirability of 
providing an indication of the voice pitch. 
Unless the acoustically handicapped are able to control the pitch 
and the volume of the voice as well as the positions of the articulators, 
their speech will sound unnatural although it may be intelligible. 
The best ways to display pitch are still uncertain. One possibility 
is by means of a wave trace perhaps below the pattern; another by a 
fixed trace, in which the line intensity is varied in proportion to the 
fundamental frequency. The latter is simpler from an apparatus 
standpoint, but in a first experimental use seemed less acceptable as 
an indication. 
There remains to be discussed potential uses for these sound pat- 
terns in various fields of specialized acoustics for purposes of analysis 
and illustration. The foregoing discussion of speech patterns indi- 
cates in a general way the possibilities as applied to phonetics, philol- 
ogy, and speech correction and development, but the patterns have 
many uses for the visual interpretation of complex waves other than 
speech. Plates 8 and 4 include a number of pictures of more or 
Jess familiar sounds that may illustrate better than speech the rela- 
tionship between what we hear and what we would see in this form . 
of visible interpretation. Sustained tones produce horizontal lines as 
in A of plate 3. The clicks of a hammer against a metal block contain 
brief spurts of energy spread over the whole frequency range, so that 
they appear as vertical lines in B of plate 3. Swinging the frequency 
of a variable oscillator up and down the scale results in the wavy line 
of C in the same figure. 
The remaining patterns of plate 3 and those of plate 4 are described 
fairly well by the brief captions. For those interested in a more 
detailed examination of the time and frequency dimensions it should 
be added that in plate 3 (A to G inclusive and I and J) and in plate 
4 (K to H inclusive) the length is approximately 9.4 seconds and the 
vertical scale includes a frequency range of zero (at the black base 
line) to 3,200 cycles per second at the top. In H of plate 8 and in the 
first four patterns of plate 4 (A to D inclusive) the length is approxi- 
mately 4 seconds and the frequency scale is zero to 7,500 cycles per 
second. 
The bird songs? pictured in plate 4 were originally selected for use 
as test material because they contain a wide variety of tone modula- 
7 These sound spectrograms were made from the Cornell Bird Song Records (Albert BR. 
Brand Bird Song Foundation, Laboratory of Ornithology, Cornell University). 
