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.

43 lines
1.6 KiB

From text to speech: The MITalk system
consonant clusters. Characterization of even the static properties of these phonetic
segments is beyond the scope of the present chapter.
11.3.3 Structure of the output parameter file
The output file consists of one complete set of control parameter values per 5 msec
of speech. The control parameters that are varied are identified in Table 11-3.
Table 11-3: Variable control parameters specified in PHONET
N Symbol Name
1 AV amplitude of voicing in dB
2 AF amplitude of frication in dB
3 AH amplitude of aspiration in dB
4 AVS amplitude of sinusoidal voicing in dB
5 FO voicing fundamental frequency in Hz
6 F1 first formant frequency in Hz
7 F2 second formant frequency in Hz
8 F3 third formant frequency in Hz
9 F4 fourth formant frequency in Hz
10 FNZ nasal zero frequency in Hz
11 Bl first formant bandwidth in Hz
12 B2 second formant bandwidth in Hz
13 B3 third formant bandwidth in Hz
14 A2 second paralle] formant amplitude in dB
15 A3 third parallel formant amplitude in dB
16 A4 fourth parallel formant amplitude in dB
17 AS fifth parallel formant amplitude in dB
18 A6 sixth parallel formant amplitude in dB
19 AB bypass path amplitude in dB
20 not currently used
11.4 Summary
PHONET differs from a number of other formant-based synthesis-by-rule
programs (e.g. Votrax, Kurzweil, Holmes, Mattingly, Rabiner, or Hertz) primarily
in terms of the total number of contextg-dependent rules that have been formulated
in order to model details of the spectra of phonetic transitions. A complete
description of these rules is given in Appendix C.
122