1 |
dpavlin |
1 |
CR1 |
2 |
|
|
-------------------------------------------------------------------------- |
3 |
|
|
Croatian male voice for the MBROLA speech synthesizer |
4 |
|
|
|
5 |
|
|
Copyright (C) 1998 by Juraj Bakran and Nikolaj Lazic, University of Zagreb |
6 |
|
|
-------------------------------------------------------------------------- |
7 |
|
|
Table of Contents |
8 |
|
|
-------------------------------------------------------------------------- |
9 |
|
|
|
10 |
|
|
1.0 Description of CR1 |
11 |
|
|
2.0 Distribution |
12 |
|
|
3.0 Installation, and tests |
13 |
|
|
4.0 Acknowledgments |
14 |
|
|
|
15 |
|
|
-------------------------------------------------------------------------- |
16 |
|
|
1.0 Description of CR1 |
17 |
|
|
-------------------------------------------------------------------------- |
18 |
|
|
|
19 |
|
|
CR1 is first version of a diphone database for Croatian. It includes 930 |
20 |
|
|
diphones. |
21 |
|
|
|
22 |
|
|
The following phoneme symbols are assumed in our diphone set: |
23 |
|
|
|
24 |
|
|
Note: This set uses the SAMPA notation for Croatian. In the representation |
25 |
|
|
of Croatian orthography shown below, |
26 |
|
|
|
27 |
|
|
* c' stands for c with an acute accent; |
28 |
|
|
* d' stands for a barred d; |
29 |
|
|
* c^ s^ z^ stand for c s z with a wedge accent (hacek). |
30 |
|
|
|
31 |
|
|
|
32 |
|
|
Vowels. There are five vowels in Croatian. Additionally, r can function as |
33 |
|
|
a vowel. |
34 |
|
|
|
35 |
|
|
SAMPA symbol Orthography Transcription Gloss |
36 |
|
|
|
37 |
|
|
i sir "sir <F> cheese |
38 |
|
|
|
39 |
|
|
e pet "pe:t <F> five |
40 |
|
|
|
41 |
|
|
a sat "sa:t <F> watch |
42 |
|
|
|
43 |
|
|
o bol "bo:l <F> pain |
44 |
|
|
|
45 |
|
|
u put "pu:t <F> way |
46 |
|
|
|
47 |
|
|
r prst "prst <F> finger |
48 |
|
|
|
49 |
|
|
Consonants. The Croatian consonant system comprises 17 obstruents and 7 |
50 |
|
|
sonorant phonemes. |
51 |
|
|
|
52 |
|
|
SAMPA symbol Orthography Transcription Gloss |
53 |
|
|
|
54 |
|
|
p pas "pas <F> dog |
55 |
|
|
|
56 |
|
|
b bas "bas <F> base |
57 |
|
|
|
58 |
|
|
t tip "ti:p <F> type |
59 |
|
|
|
60 |
|
|
d dan "da:n <F> day |
61 |
|
|
|
62 |
|
|
k kap "ka:p <F> drop |
63 |
|
|
|
64 |
|
|
g gol "go:l <F> naked |
65 |
|
|
|
66 |
|
|
ts cura "tsura <F> girl |
67 |
|
|
|
68 |
|
|
tS c^ar "tSa:r <F> thrill |
69 |
|
|
|
70 |
|
|
dZ dz^ep "dZep <F> pocket |
71 |
|
|
|
72 |
|
|
tS' noc' "no:tS'<F> night |
73 |
|
|
|
74 |
|
|
dZ' d'ak "dZ'a:k <F> pupil |
75 |
|
|
|
76 |
|
|
f fonetika fo"netika<R> phonetics |
77 |
|
|
|
78 |
|
|
s sat "sa:t <F> hour |
79 |
|
|
|
80 |
|
|
z zid "zi:d <F> wall |
81 |
|
|
|
82 |
|
|
S s^ef "Se:f <F> chief |
83 |
|
|
|
84 |
|
|
Z z^ena "Zena <R> woman |
85 |
|
|
|
86 |
|
|
x hod "xo:d <F> walking |
87 |
|
|
|
88 |
|
|
m most "mo:st <F> bridge |
89 |
|
|
|
90 |
|
|
n noga "noga <R> leg |
91 |
|
|
|
92 |
|
|
J njiva "Jiva <F> field |
93 |
|
|
|
94 |
|
|
l lud "lu:d <F> mad |
95 |
|
|
|
96 |
|
|
L ljut "Lu:t <F> bitter |
97 |
|
|
|
98 |
|
|
r ruka "ru:ka <R> hand |
99 |
|
|
|
100 |
|
|
v vid "vi:d <F> sight |
101 |
|
|
|
102 |
|
|
j ja "ja: <F> I |
103 |
|
|
|
104 |
|
|
-------------------------------------------------------------------------- |
105 |
|
|
2.0 Distribution |
106 |
|
|
-------------------------------------------------------------------------- |
107 |
|
|
This distribution of cr1.zip contains the following files : |
108 |
|
|
|
109 |
|
|
cr1 - Croatian male voice |
110 |
|
|
cr1.txt - this file |
111 |
|
|
|
112 |
|
|
and a number of example .PHO files : |
113 |
|
|
|
114 |
|
|
brazilac.pho |
115 |
|
|
kantica.pho |
116 |
|
|
zalostan.pho |
117 |
|
|
naslov.pho |
118 |
|
|
example.pho |
119 |
|
|
|
120 |
|
|
Please consult the MBROLA project homepage : |
121 |
|
|
|
122 |
|
|
http://tcts.fpms.ac.be/synthesis/mbrola.html |
123 |
|
|
|
124 |
|
|
Registered users will automatically be notified of the availability |
125 |
|
|
of new databases. To freely register, simply send an email to |
126 |
|
|
mbrola-interest-request@tcts.fpms.ac.be with the word 'subscribe' |
127 |
|
|
in the message title. |
128 |
|
|
|
129 |
|
|
-------------------------------------------------------------------------- |
130 |
|
|
3.0 Installation and Tests |
131 |
|
|
-------------------------------------------------------------------------- |
132 |
|
|
|
133 |
|
|
If you have not copied the MBROLA software yet, please consult |
134 |
|
|
the MBROLA project homepage and get it. Also refer to the readme.txt file |
135 |
|
|
provided with the mbrola software for using it. |
136 |
|
|
|
137 |
|
|
Copy cr1.zip into the mbrola directory and unzip it. |
138 |
|
|
|
139 |
|
|
Try |
140 |
|
|
|
141 |
|
|
mbrola cr1 zalostan.pho zalostan.wav |
142 |
|
|
|
143 |
|
|
to create a sound file for the word /Zalostan/ (sad). Listen to it with |
144 |
|
|
your favorite sound editor, and try the other command files (*.pho) to |
145 |
|
|
have a better idea of the quality of speech that can be synthesized with |
146 |
|
|
MBROLA and the cr1 database. |
147 |
|
|
|
148 |
|
|
In the creation of new *.pho files, there is no need to worry about |
149 |
|
|
allophones. They are included in the database by allowing natural |
150 |
|
|
assimilation in the course of pronunciation. You need simply write the |
151 |
|
|
phonemic (SAMPA) transcription in the *.pho file. |
152 |
|
|
|
153 |
|
|
-------------------------------------------------------------------------- |
154 |
|
|
4.0 Acknowledgments |
155 |
|
|
-------------------------------------------------------------------------- |
156 |
|
|
|
157 |
|
|
The cr1 database was developed in Department of Phonetics at University |
158 |
|
|
of Zagreb. |
159 |
|
|
|
160 |
|
|
We would like to thank Vincent Pagel for his efforts in converting our |
161 |
|
|
diphone database to the MBROLA format and for all suggestions during the |
162 |
|
|
construction and transfer of the database. Also, we would like to thank |
163 |
|
|
to all students of phonetics who took part in the construction of the |
164 |
|
|
database. |
165 |
|
|
|
166 |
|
|
|
167 |
|
|
July 1, 1998. |
168 |
|
|
|
169 |
|
|
Juraj Bakran and Nikolaj Lazic |
170 |
|
|
|
171 |
|
|
For more information contact: |
172 |
|
|
|
173 |
|
|
juraj.bakran@ffzg.hr or |
174 |
|
|
nikolaj.lazic@ffzg.hr |
175 |
|
|
|
176 |
|
|
or write to: |
177 |
|
|
Department of Phonetics, I. Lucica 3, 10000 Zagreb, CROATIA |
178 |
|
|
-------------------------------------------------------------------------- |