Page 123 - The Master Handbook Of Acoustics

98 CHAPTER FIVE

Noise amplitude
control signal

Random-
“Unvoiced” number
generator
Digital Digital
formant to
filter analog

Pitch
“Voiced” pulse
generator
Filter
control
signal
Pitch Voice
frequency intensity
control control
signal signal
FIGURE 5-7
A digital system for synthesizing speech. Note the similarity to the models of Figs. 5-3, 5-4, and 5-5.

Other applications of digital speech synthesis include voice-recog-
nition systems for “hands-free” typing and computer programs for
recording that automatically adjust the intonation of a vocalist who
might otherwise be singing sharp or flat.

Directionality of Speech
Speech sounds do not have the same strength in all directions. This is
due primarily to the sound shadow cast by the head. A question arises
as to just how such directionality can be measured. Should a sound
source be placed in the mouth? Well, the mouth itself is a continuous
source of speech sounds, so why not use these speech sounds for the
measurement of directionality? That is what Kuttruff has done with
2
the results shown in Fig. 5-8. Because speech sounds are highly vari-
able and extremely complex, careful averaging is necessary to give an
accurate measure of directional effects.
The horizontal directional effects, shown in Fig. 5-8A, show only a
modest directional effect of about 5 dB in the 125- to 250-Hz band.