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
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
|
LIBDUMBTTS
Libdumbtts is library extending capabilities of "dumb synthesizers",
which can say only input text including characters of language alphabet.
This library enriches:
1) translated key names
2) saying single characters; it recognizes all characters
between 0x20 and 0x17f.
In case of letters, it is possible to recognize capital letters
by adding prefix 'capital' or by using function returning information
if said letter is capital (helpful when you're using sound icon
prefix or pitch change).
Accented letters are pronounced by connecting character and name of accent
(ie. "o acute" for '')
3) converting input unicode text to one-byte synthesizer encoding.
Pronunciation of all letters is configurable. This library implements
also simple transliteration method of cyryllic (for characters range:
0x400 - 0x47f).
Library is projected to be used by Ivona synthesizer and
speech-dispatcher. On the other side, it should be as flexible
as it can be to generate output texts for every other synthesizer.
METHODS:
struct dumbtts_conf *dumbtts_TTSInit(char *lang);
Initializes dumbtts_conf structure and reads config file.
Configuration files should be placed in /usr/share/dumbtts/xx.conf,
where xx is two-character code of language. After reading this file,
function is trying to search /etc/dumbtts/xx.conf and then
~/.dumbtts_xx.conf file and corrects elements read before.
It returns allocated memory, which cannot be freed by free() function!
void dumbtts_TTSFree(struct dumbtts_conf *conf);
It frees memory allocated by dumbtts_TTSConf.
int dumbtts_WCharString(
struct dumbtts_conf *conf,
wchar_t wc,
char *outbuf,
int len,
int capMode,
int *isCap);
int dumbtts_CharString(
struct dumbtts_conf *conf,
char *ch,
char *outbuf,
int len,
int capMode,
int *isCap);
It pronounces character (in case of CharString, first character
of input). Output text is allocated in area pointed by parameter
'outbuf', sized at least 'len' of bytes. 'capMode' parameter
indicates way of announcing capital letters. It can be:
0 - no capital letters announcing
1 - announcing capital letter by application
2 - preceding capital letter by word 'capital'
(or language-specific equivalent)
At this moment, values 1 and 0 works in the same way.
isCap parameter should include pointer to integer variable or NULL.
If not NULL, value of variable will be set to 1 in case of capital
letter and to 0 in case of other character.
Returns:
If conversion succeeds, function returns 0. If outbuf parameter
was NULL or given size of pointed area is to small to fit
converted text, it returns minimal size of area. In case of error,
it returns negative value.
int dumbtts_KeyString(
struct dumbtts_conf *conf,
char *str,
char *outbuf,
int len,
int capMode,
int *isCap);
Converts name of key (eg. "space", "underscore") included in 'str'
to text translated to local language and pronouncable by synthesizer.
Meaning of parameters and returning value - the same as for CharString.
int dumbtts_GetString(
struct dumbtts_conf *conf,
char *str,
char *outbuf,
int len,
int punctMode,
char *punctChars,
char *punctLeave);
Converts given input to form pronouncable by synthesizer.
punctMode parameter can be set by three possible values :
0 - punctuation characters are ignored
1 - punctuation is limited to characters given
in punctChars parameter
2 - all punctuation characters are spoken.
punctLeave parameter is input containing characters, which
have to be left unchanged in output, regardless the punctMode
value - they are basic punctuation characters controlling
pronounciation. For Ivona, value of input should be ".,!?;:".
Meaning of rest parameters and returning value - the same as
for CharString.
CONFIGURATION FILES
File is divided into sections. Name of each is enclosed in square brackets.
During parsing, first four letters of name are read. Empty lines are ignored.
Comments are preceded by "//" string. Files have to be in UTF-8.
For examples - see pl.conf file.
Meaning of sections:
SECTION [clist]
Includes declaration of all letters recognized by synthesizer by :
number character
where "number" jest representation of character in encoding used by
synthesizer and "character" is single small letter.
SECTION [spell]
Contains pairs character-name, used when pronouncing single characters.
Character can be encoded in #number notation (decimal) or in #xnumber
(hexadecimal). Read number is adequate to character code in UNICODE.
Example :
y igrek
SECTION [pronunciation]
Contains pronounciation tables for each characters used when reading texts.
SECTION [diacritics]
Contains names of diacriticts and some reserved words.
ill illegal // illegal symbol etc.
let letter
cap capital
caplet capital letter // if not same as 'cap' 'let'
lig ligature
caplig capital ligature // if not same as 'cap' 'lig'
sym symbol
ced cedilla
bre breve
dot dot above
dia diaeresis // umlaut
acu acute
gra grave
ac2 double acute // hungarumlaut
cir circumflex
mac macron
rin ring // above
str stroke
ogo ogonek
til tilde
car caron // hacek
SECTION [keys]
Includes pairs name - translation used when reading keys.
SECTION [cyrillic]
Contains pairs characters - transliteration used when reading cyryllic.
"Characters" are single single letter or two small letters in cyryllic,
transliteration can be empty.
SECTION [abbreviations]
Contains pairs abbreviation - pronunciation. Abbreviation may be
followed by dot (to prevent pass dot to synthesizer).
Please send every suggestion and bugfix requests to:
Bohdan R. Rau <ethanak@polip.com>
|