1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
|
This is an example of the intermediate file format used in the
program. The format is usually called the ``.spn'' file format
after the file extention used. I don't know what it stands for.
This data would be sent from the perl part of the program to the
synth sub-program. A typical ``.spn'' file might look like the
following. Note there are no blank lines. You wont actually get
to see an spn file unless you hack the program a bit, but if you
want to use, for example, the synth program separately from the high
level, then you need to know what the format is.
----------start of data------------------------
# 50 (0,120)
aa 120 (0,100) (30,130)
b 73
a 46
d 48 (0,100)
ee 92 (30,130) (80,90)
n 59 (99,80)
# 1200 (99,80)
----------end of data------------------------
The input format is a sequence of lines of the form <phoneme symbol>
<duration> (<pitch target>...).
The set of phoneme symbols is specified in the file ``t2s''.
Durations are given in milliseconds.
A pitch target pair has the format (percentage,frequency) where
frequency is in hertz, and percentage is an integer in the range 0-99
inclusive. The percentage specifies the proportion of the phoneme
duration that has elapsed at the time when the target is achieved.
Note that the format of the file needs to be carefully adhered to.
There are, for example, no spaces allowed within the bracketed pairs
that specify the pitch targets.
The initial and final silences (#'s) are obligatory, and they must be
accompanied by targets with percentage values of 0 and 99
respectively. (That is, there must be targets specifying the pitch at
the beginning and end of the utterance.) The initial and final
durations are allowed to vary (but at the moment are not respected).
|
