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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
|
# Copyright 2001,2002,2003,2004,2005 Sun Microsystems, Inc.
# All Rights Reserved. Use is subject to license terms.
#
# See the file "license.terms" for information on usage and
# redistribution of this file, and for a DISCLAIMER OF ALL
# WARRANTIES.
#
Change history for FreeTTS.
--------------------------
March, 9, 2009 Release 1.2.2:
- Introduced an AliasDiphone class as a subclass of Diphone, which
has a name and an original diphone; most API calls are just
forwarded to the original diphone. (Marc Schroeder)
- In text db files, introduced a new type of line, "ALIAS new
old", where "old" is the name of an existing diphone (must have
been defined higher-up in the text db file using the traditional
"DIPHONE...." mechanism), and "new" is the new name under which
the diphone is to be known additionally.
DiphoneUnitDatabase.parseAndAdd () has been adapted accordingly.
(Marc Schroeder)
- In binary db files, introduced a new magic code identifying an
alias diphone entry. Diphone.loadBinary() has been adapted
accordingly. (Marc Schroeder)
- Fixed "Class-Path" issue in manifest created by FestVoxToFreeTTS.
(Willie Walker)
- Fixed internal class loader problems (Dirk Schnelle-Walka)
- using java logging library instead of system.out.println
- separation into a core jar and a jsapi1.0 dependent jar
(Dirk Schnelle-Walka)
- Bugfixes in JavaClipAudioPlayer ((Dirk Schnelle-Walka):
Catching runtime exceptions if the line is not available
writing directly to the output line instead of buffering the data
- some code cleanup (Dirk Schnelle-Walka)
March 10, 2005 Release 1.2.1
- Add support for metrics important to screen readers
- Add drain delay to clip audio player - the lack of
it was causing the end of speech to be cut on JDS/Linux
February 06, 2005 Release 1.2 (final)
- Hang with JDK1.5 (bug 1019337)
- Added support to build w/o requiring JSAPI
- Only build tests on demand (more freedom from JUnit)
- Added build targets to create standalone emacspeak app
- Added ability for emacspeak demo to strip DECTalk commands from text
- Added speaking rate change support to emacspeak demo
- Allowed emacspeak text to omit '{' and '}' if text has no spaces
- Changed state name expansion logic (bug 1101229)
- Added pronunciations for British Sterling and "not" sign (bug 1100888)
- Modified missing speech.properties message (rfe 1041437 and bug 1041296)
- Added pronunciations for Linux terms (bug 1006630)
- Added JUnit to prerequisites list (bug 975557)
- Fixed -dump{Multi}Audio problem in freetts app (bug 986503).
- Added importing tools specifically for CMU ARCTIC Voices.
- Fixed problem in mbrola phonem remapping (bug 1013781)
- In the Voice class, removed the call to AudioPlayer.drain()
if the audio is cancelled, since it will block until the next audio
is available.
- Fixed problem in FreeTTSSynthesizer where in some cases waiting
threads are not notified when a queue item is cancelled.
- Made the JavaStreamingAudioPlayer.cancel() method unsynchronized
so that audio output can be stopped immediately.
- Added rudimentary internationalization to Festvox import.
This imports international (non-US-English) cluster unit voices,
but obviously still lacks a lexicon and other NLP components for
new languages.
June 15, 2004 Release 1.2beta2
- Added experimental CMUArcticVoice that may (or may not) work
with the CMU ARCTIC databases.
- Added CMUTimeVoice, which has noop post lexical processing
to support the FestVox limited domain time voice scheme.
Modified FestVoxToFreeTTS to use this voice in the template.
- Restructured cluster unit voice for simple extensions.
- Added com.sun.speech.freetts.lexicon.userAddenda property
- Added FreeTTS pronunciation to addenda
- Fix problem where the space character in the path would cause
issues with finding voice jars (bug 938225)
- Added backwards compatibility for Gnome Speech API
- Fix to kill audio output thread on deallocate from JSAPI (bug 922579)
- Added "@" and "?" to list of known characters for German.
- Fixed problem where some jar files in the FreeTTS directory
would cause issues with discovering voices.
- Added defaultAudioPlayer property to voice (RFE 922580).
December 22, 2003 Release 1.2beta
- Many documentation updates to prep for beta release.
November 22, 2003 Release 1.2beta
- Abandon GNUMake in favor of Ant.
- Many documentation updates.
- Add capability to set the AudioPlayer on a JSAPI Voice.
September 3, 2003 Release 1.2beta
- Modified MBROLA output to work on Solaris and Mac
August 15, 2003 Release 1.2beta
- Added support to import voices from FestVox into FreeTTS
- Extended voice packaging scheme to simplify management and addition
of new voices.
- VoiceManager class provices API to access voices.
- Dynamic detection of voice jarfiles is supported.
Jun 5, 2003 Release 1.2beta
- Updated FreeTTS version message
Jun 4, 2003 Release 1.2beta
- Imported changes from Flite 1.2:
- Regenerated cmulex_addenda, cart_dump
- Reads year "1200" as "twelve hundred" rather than "twelve oh
zero"
- TimeUtils reads 00:01 as "just after midnight" rather than
"just after midnight, in the morning".
- Added "daughter1" to be an alias for "daughter"
- Minor bug fixes in TokenToWords and ClusterUnitSelector
- Regenerated test data to compare FreeTTS against Flite 1.2
- Wrote script gen_flite_data.sh to help automate updates to test
data.
May 13, 2003 Release 1.1.2
- Fixed bug in TokenToWords that causes abbreviations like "OK"
and "HI" to be spoken as a US state rather than just "ok" and "hi".
- Fixed race condition in JavaStreamingAudioPlayer that would
cause output to not be totally cancelled.
- Refactored all System.getProperty calls to call a wrapper
utility called Utilities.getPropery. This method catches a
SecurityException and deals with it properly, so that it can be
called from an applet.
- Fixed problem in the JSAPI implementation where the audio player
was not being closed when the synthesizer was deallocated. This
would cause problems for certain types of AudioPlayers such as
SingleFileAudioPlayer that require a 'close' before they flush
their accumulated wave data.
September 6, 2002 Release 1.1.1
- Fixed index.html so that it looks for the RELEASE_NOTES in
the proper spot.
- Added RELEASE_NOTES to the deploy_doc Make target.
- Fixed all run.bat files to run properly under Win32 systems.
- Fixed race condition in com.sun.speech.engine.EngineMonitor that
would cause a NullPointerException in the Player demo on
occasion.
- Added emacspeak deploy options
- Added non-JSAPI version of Emacspeak speech server
- Fixed non-JSAPI Emacspeak demo crashing problem by using the
JavaStreamingAudioPlayer (instead of ClipPlayer) as the default.
- Added RawFileAudioPlayer to write raw audio bytes to a binary file.
- Changed all JDK_DIR references to JAVA_HOME
- Fixed problem in parsing certain dates (MonthRange)
- Fixed problem with some number parsing
- Removed hardcoded array size of LPCResult.residualToFloatMap
- Fixed problem where output thread was not being shut down when a
voice was closed.
- Added TokenGame test which exercises some tricky tokenizing
- Screen shots of Player in demo/JSAPI/Player/readme.html now of
current version.
- parameterized utterance tests to allow easy addition of more
tests
- demo/freetts/ClientServer batch files specifies host and port
- Added contributors agreement: ContributorsAgreement.pdf
- Added handout: FreeTTS_handout.pdf FreeTTS_handout.sxw
- Added 'How to contribute' section to the web page
- Refactored identical demo code in the JSAPI and freetts branches
so that they share the same code (put in demo/util directory).
- Added "exit" command to Emacspeak. This will close the socket
connection.
- Added JSAPI/Web Start clock demo to show how to use FreeTTS/JSAPI
with Java Web Start.
- Added support for MBROLA Voices, as well as JSAPI support for
MBROLA Voices. Specifically, the latter refers to check for
the MBROLA binaries and databse for an MBROLA voice. If they
are found, such a voice is consider 'valid'.
- Fixed the following bug: A word followed by a breaking
punctuation (:?!), followed by space and a single word,
e.g., "RE: Frank", causes the word following the breaking
punctuation (i.e., "Frank") not to be synthesized.
- Fixed the following bug: "From:" or "From" causes the system to
hang.
- Added MBROLA support to all appropriate demos.
January 30, 2002 Release 1.1
- Created index.php to automatically put news on freetts front page
- Added a very small, time-domain specific lexicon
- Updated voice to allow easier switching betwen different
lexicons
- Adapted cluster unit database layout to conform to flite 1.1
format
- Adapted cluster unit selection to conform to new database
layout
- Upgraded 16khz diphone database with Flite 1.1 data.
- Added support for duration shrinking/stretching at the segment
level.
- Added support for pitch shift adjustments at the segment level.
- More sophisticated phrasing CARTS.
- PostLexicalAnalyzer now deals with "'ve" and "'ll"
- Some refactoring of the FeatureProcessors
- A number of new Feature processors have been added.
- Added ASCII file support for Pronounceable finite state machines
- Expanded tokenizer. New features include:
1) processing abbreviations (e.g. U.S.A.)
2) 7-digit phone numbers,
3) partial phone numbers
4) time (e.g. 12:35)
5) digits w/ dashes (e.g. 999-999-999)
6) Roman numerals (e.g. IV)
7) St Andrew's St.
8) US Money (e.g. $14.53)
9) "80s" or "70s"
10) fractions (e.g. 3/4)
11) US states (e.g. NH, MA)
- Improved rules for generating binary forms of the databases
- Specifies 'Locale.US' (instead of null) when constructing a
SynthesizerModeDesc in the demo code
- Output of regression tests are more consistent
- Workaround for the Linux sound bug reduces JVM crashes.
January 8, 2002 Release V1.0.5
- Fixed Locale.US problem in JSAPI demos.
January 8, 2002 Release V1.0.4
- Fixed speech.properties path message in demos.
- Added SourceForge news to docs front page.
December 21, 2001 Release V1.0.3
- Added .bat files for win32 demos
December 20, 2001 Release V1.0.2
- Added jsapi.exe - win32 jsapi installer
December 20, 2001 Release V1.0.1
- Added jsapi.sh - unix jsapi installer
December 14, 2001 Release V1.0
- First public release
November 7, 2001 Release V0.9
- Internal Test Release
|