You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
31 lines
1.2 KiB
31 lines
1.2 KiB
From text to speech: The MITalk system
|
|
|
|
INPUT OUTPUT
|
|
R4 R1
|
|
|
|
CASCADE
|
|
|
|
INPUT OUTPUT
|
|
|
|
PARALLEL
|
|
Figure 12-3: Parallel and cascade simulation of the vocal tract transfer function
|
|
|
|
dividual amplitude controls for each formant. The disadvantage is that one still
|
|
needs a parallel formant configuration for the generation of fricatives and plosive
|
|
bursts -- the vocal tract transfer function cannot be modeled adequately when the
|
|
sound source is above the larynx, so that cascade synthesizers are generally more
|
|
complex in overall structure.
|
|
|
|
A second advantage of the cascade configuration is that it is a more accurate
|
|
model of the vocal tract transfer function during the production of nonnasal
|
|
sonorants (Flanagan, 1957). It will be shown that the transfer functions of certain
|
|
vowels cannot be modeled very well by a parallel formant synthesizer. Although
|
|
not optimal, a parallel synthesizer is particularly useful for generating stimuli that
|
|
violate the normal amplitude relationships between formants, or if one wishes to
|
|
generate, e.g., single-formant patterns.
|
|
|
|
The software simulation to be described has been programmed for normal use
|
|
as a hybrid cascade/parallel synthesizer (Figure 12-4a), or alternatively for special-
|
|
|
|
126
|