Debian Codegraph @

NEWS

raw
1 Version 0.12.1
2 * Fix spd-say -k ' '
3 * Fix output getting stuck in some rare condition triggered in Orca preference dialog
4 * Fix taking character punctuation levels from NVDA files.
5 * espeak: Fix translated pronounciation.
6 * generic: Fix getting stuck on speaking nothing.
7 * generic: Add GenericDefaultCharset parameter
8 * generic: Fix random garbage on non-text utterances
9 * Fix Czech documentation encoding.
10
11 Version 0.12
12 * Add openjtalk module.
13 * Add socket activation.
14 * Add language and variant parameters to "list voice" command.
15 * Add libspeechd-module library for making it simpler to create external spd modules.
16 * libspeechd: Simplify buffer management.
17 * Fix use of semaphores on MacOS.
18 * Fix splitting long UTF-8 sequences.
19 * Update CLDR to version 45, symbols from orca 45.2, and symbols from NVDA.
20 * Add spd_fd function to C api.
21 * Detect module failures from generic module.
22 * Make the fallback espeak-ng and dummy modules hardcoded.
23 * Better detect generic module failures to disable them.
24 * pulse: Use asynchronous API to avoid buffer underruns.
25 * generic: Make stripping punctuation use locale charset.
26 * audio: Fix logging from audio modules in server-side audio.
27 * Sort Baratinoo engine higher.
28 * espeak-ng-mbrola: Fix mbrola voices with rate different from 22KHz
29 * Add a run-spd-say script and make run-speechd and run-spd-say able to talk directly.
30 * server: fix daemon mode on macOS with libao.
31 * Add initial pipewire support.
32 * Also support loading symbols from home directory
33
34 Version 0.11.5
35 * Update CLDR to version 43 and symbols from NVDA.
36 * Fix parsing and emitting space character.
37 * Reduce espeak buffer size to avoid ssml issues.
38 * Add --disable-doc.
39 * Fix spd-conf not being able to set the default module.
40 * Fix pausing without index.
41 * ibmtts: Fix parameter reset.
42 * ibmtts: Disable using the module punctuation by default.
43
44 Version 0.11.4
45 * Update CLDR to version 42 and symbols from NVDA.
46 * Fix audio plugin loading with dlopen.
47 * Fix atomicity of getting reply in threaded mode.
48
49 Version 0.11.3
50 * Fix back DefaultModule configuration.
51 * pico: Avoid falling to english when passed a bogus voice name.
52 * espeak: Fix setting voice type.
53
54 Version 0.11.2
55 * Fix loading xx-yy locales.
56 * Various memory leaks fixes.
57 * Add mimic3 configuration file.
58 * pico: Fix setting language vs voice.
59 * Make sure that modules report a list of voices.
60 * Update CLDR to version 41, symbols from NVDA and orca.
61 * Allow building without ltdl.
62 * Re-enable SSML in espeak-ng-mbrola module.
63
64 Version 0.11.1
65 * Add SPEECHD_PLUGIN_DIR environment variable.
66 * Fix listing voices of the default module.
67
68 Version 0.11
69 * Support playing audio through the server.
70 * Simplify modules thanks to this.
71 * modules: Add support for loading from user's .local/libexec/speech-dispatcher.
72 * symbols: Process symbols.dic before emojis.dic.
73 * symbols: Enable speechd symbols processing by default.
74 * modules: Moved speech dispatcher modules to /usr/libexec/speech-dispatcher-modules
75 (but keep compatibility with old /usr/lib/speech-dispatcher-modules path).
76 * espeak-ng: Add support for mbrola voices.
77 * ibmtts/voxin: Improve language selection.
78 * ibmtts/voxin: Add option to disable punctuation commands.
79 * ibmtts: Fix library load.
80 * ibmtts: Do not build by default on other systems than 32bit PC.
81 * mary: Add auto-detection.
82 * mary: Add newer voices.
83 * mary: Add volume, pitch, and rate support.
84 * ivona: Add auto-detection.
85 * festival: Strip head silence.
86 * generic: Add DefaultVoice option.
87 * es_ES: Add some gender neutral rules.
88 * Add SPEECHD_CMD environment variable.
89 * modules: Rewrite main functions with BSD licence, to let proprietary modules easily reuse this as a basis.
90 * modules: Add skeletons ready for use as a basis for new modules.
91 * Add script to run speechd from the build tree.
92 * Update CLDR to version 39, symbols from NVDA and orca.
93 * Add Esperanto translation.
94 * Sort modules by quality, let the best quality module be the default.
95
96 Version 0.10.2
97 * generic: Add support for sound icons
98 * Add French, Norwegian Bokmål, Portuguese, Brazilian Portuguese, and Russian translations.
99 * Add voxin module variant.
100 * Factorize espeak module with espeak-ng module.
101 * Make ibmtts and voxin use the common queue helpers.
102 * Fix python bindings against python 3.9.
103 * Reference the golang bindings.
104 * Add spd_get_client_id.
105 * Make key, char and sound_icon commands return message id.
106 * Expose modulebindir in speech-dispatcher.pc.
107 * Remove generic modules for synths that have a non-generic module.
108
109 Version 0.10
110 * Add punctuation level "most".
111 * Rework thread synchronization.
112 * Move espeak/espeak-ng audio queuing to shared file, rework it.
113 * Make baratinoo use this audio queueing, enable marks, and support
114 responsiveness parameter.
115 * Add support for Baratinoo engine 8.6
116 * Show mark progression in spd-say.
117 * Add --character option to spd-say.
118 * Put country/region in language instead of variant.
119 * pico, generic: fix setting voice by language.
120 * generic: Add fallback voices for generic languages.
121 * symbols: Add support for SSML.
122 * symbols: Add SymbolPreprocFile directive to load several symbol files.
123 * symbols: Replace DefaultSymbolsPreproc directive with SymbolsPreproc directive
124 to have more fine-grained control over server preprocessing.
125 * symbols: Import emojis and unicode font variants support from NVDA and Unicode
126 CLDR and UnicodeData.
127 * symbols: Import some symbols from Orca.
128 * symbols: Add rules to drop gender-neutral forms.
129 * symbols: Add support for group references.
130 * Disable Mary-TTS module by default.
131 * Replace AudioPulseServer option with AudioPulseDevice.
132 * Fix default pulse latency to 10ms to avoid pulseaudio distorted output.
133 * Made client isolation stronger for Snap/Flatpak support.
134
135 Version 0.9.1
136 * Add module for the non-free IBM TTS (voxin) speech synthesis.
137 * Extend licence to later versions of GPL and LGPL.
138 * Update mailing list address to savannah.
139 * Make generic modules fallback to existing voices.
140
141 Version 0.9
142 * Add modules for non-free Baratinoo (VoxyGen) and Kali speech syntheses.
143 * Add configuration file for the Mary-TTS system.
144 * Add configuration file for espeak-ng + mbrola.
145 * Set the pulse client name when using the generic module with paplay.
146 * espeak-*-mbrola-generic: Update voice list.
147 * Auto-detect module availability.
148 * Make generic module provide voice list.
149 * Add systemd service file.
150
151 Version 0.8.8
152 * Add German translation, thanks to Chris Leick for the patch
153 * Fix some spelling mistakes, thanks to Paul Gevers for the patch
154 * Some slight code improvements
155
156 Version 0.8.7
157 * Further fixes to spd-conf, which should now work properly.
158 * Split the espeak-ng driver code into its own source file.
159 * Add a work-around to the espeak-ng driver to account for spaces in voice
160 names which recently appeared in espeak-ng git master. This will properly
161 be fixed in 0.9.
162 * Voice names are not forced to lower case, due to espeak-ng git master now
163 having multi-case voice names.
164 * Fix stripped audio output from the flite module, thanks to Samuel Thibault.
165 * Further code and build improvements.
166
167 Version 0.8.6
168 * Various internal code improvements.
169 * Fix more compiler warnings.
170 * Python bug fixes with thanks to Sebastian Humenda.
171
172 Version 0.8.5
173 * More unused code removal.
174 * Fix more compiler warnings.
175 * Use GLib main loop for the main server thread.
176 * Implement a shutdown timer in the server, which activates after 5 seconds
177 with no clients connected.
178 * Add support for espeak-ng.
179 * Configuration documentation for the ibmtts module.
180 * Removal of unused configuration options from the ibmtts module.
181 * Add command-line argument to allow for custom modules location
182
183 Version 0.8.4
184 * Updated documentation for required dependencies and where to find them.
185 * Removed unused code.
186 * Fixed compiler and GLib warnings.
187 * Cleanup header definitions and inclusions.
188 * Enabled silent rules by default
189 * Fix language identification references.
190
191 Version 0.8.3
192 * Add API methods to get language, rate, pitch, and volume.
193 * A lot of code cleanup, and compatibility improvements.
194 * Removed all references to GNOME Speech, since it has long since been
195 deprecated.
196 * Fix some inconsistancy in the SSIP API for voice type.
197 * The SET VOICE SSIP command is now deprecated, and will be removed in 0.9.
198 * The C library API now provides macro definitions for major, minor, and micro
199 versions in libspeechd_versions.h.
200 * The libsndfile library is now a mandetory dependency to improve the user
201 experience around sound icons.
202 * Fix a possible crash in the festival driver.
203 * Add a configuration option to the espeak driver to show voice variants in the
204 voice list. This will remain until a proper variants retrieval API is added
205 for compatible synthesizers.
206
207 Version 0.8.2
208 * Add convenience methods to the libspeech API to free module list and voice
209 data structures.
210 * Add method to the libspeechd API to get the current output module, and
211 update the documentation accordingly.
212 * The API is now licensed under the GNU Lesser General Public License v2.1
213 or later.
214 * The spdconf configuration utility is now translatable.
215 * Fixed a bug where speech-dispatcher would fail to start if the user
216 configuration directory existed but did not contain a config file.
217 * Install the spdconf desktop file.
218
219 Version 0.8.1
220 * User dictionaries support added to the IBMTTS driver
221 * Added a pico configuration file for use with the generic driver
222 * Better support for multi-arch enabled distros to facilitate the use of the
223 i386 only IBMTTS driver being easily installable on an amd64 system
224 * Bug fixes, and documentation cleanup
225
226 Version 0.8
227 * Python 3 compatibility of the Python bindings
228 * User configuration, logs and runtime files are now stored in
229 directories according to the XDG specification
230 * Internationalization of spd-say and translations into Hungarian
231 and Czech languages
232 * Espeak output can now use libsonic for faster speech
233 * Pico output module
234 * Lots of bugfixes, cleanups and fine-tunnings
235
236 Version 0.7.1
237 * Easy configuration of the client communication method using a
238 single environment variable SPEECHD_ADDRESS.
239 * Advanced autospawn -- server is autostarted only when its
240 communication channel matches the communication channels requested
241 by the client. Local server is not started when client attempts to
242 connect to a remote server.
243 * Both libspeechd and python library now report a detailed error
244 information in case of connection failure (why has connect failed,
245 why was it not possible to autostart the server, what has the
246 server reported etc.) Clients can directly show this information
247 to the user and the user doesn't have to search them in logfiles.
248 * Improvements in memory usage
249 * Cleanup of priorities of messages in logging output
250 * Various bugfixes, cleanups and fine-tunnings
251
252 Version 0.7
253 * Speech Dispatcher uses UNIX style sockets as default means of communication,
254 thus avoiding the necessity to choose a numeric port and greatly easying
255 session integration
256 * Autospawn -- server is started automatically when a client requests it
257 It can be forbidden in the appropriate server configuration file
258 (thanks to Luke Yelavich)
259 * Pulse Audio output reworked and fixed (thanks to Bill Cox and Rui Batista)
260 * Dispatcher runs as user service (not system service) by default
261 and doesn't require the previous presence of ~/.speech-dispatcher
262 directory
263 * All logging is now managed centrally, not by separate options
264 (thanks to Luke Yelavich)
265 * Graceful audio fallback (e.g. if pulse is not working, use Alsa...)
266 (thanks to Luke Yelavich)
267 * Audio output to different soundsystems through the
268 libao library (thanks to Marco Skambraks)
269 * Various bugfixes and fine-tunnings
270 * Updated documentation
271
272 Version 0.6.7
273 * Central settings for preferred audio output method
274 * Dummy output module gives a pre-recorded message whenever
275 no other output module works
276 * Possibility to switch on verbose logging over SSIP
277 for easy debugging and bug-reporting from client applications
278 * 'spd-conf' configuration, diagnostics and troubleshooting tool
279 * Volume settings in Pulse Audio
280 * Pulseaudio output now keeps connection to daemon open.
281 * Punctuation mechanism in IBM TTS is now configurable
282 * New generic output modules for Espeak with Mbrola
283 and for Cepstral Swift.
284 * Bugfixes
285
286 Version 0.6.6
287 * Bugfixes (ALSA audio output, libspeechd connection
288 failure handling and others)
289
290 Version 0.6.5
291 * Pulse Audio output module (thanks to Gilles Casse)
292 * Speech Dispatcher is now adapted for easy setup and using
293 under ordinary system user accounts
294 * Bugfixes
295
296 Version 0.6.4
297 * Native output module for eSpeak (thanks to Lukas Loehrer)
298 * Support for punctuation settings in the Generic output module
299 (thanks to Steve Holmes)
300 * Default audio output for all modules switched to ALSA
301 * Support for listing output modules and voices supported by
302 a synthesizer
303 * Bug fixes (SMP related bugs, bugs in libspeechd and others)
304
305 Version 0.6.2
306 * Bug fixes.
307 * Generic output module for eSpeak includes definitions for all
308 known languages.
309 * Python interface now supports callbacks.
310
311 Version 0.6.1
312 * Bug fixes
313 * Generic output module support for the eSpeak synthesizer (free, GPL)
314 * Output module for Cicero (french TTS, GPL but requires mbrola)
315 * Output module for IBM TTS (non-free)
316 * Revision and stabilization of the Python interface
317
318 Version 0.6
319 * ALSA and experimental NAS sound output supported (apart from OSS)
320 * SSIP implementation now supports events notification and index marking
321 * Improved documentation
322 * spd-say client functionality expanded (stopping, client name setting)
323 * Better performance
324
325 Version 0.5
326 * SSIP now supports SSML (Speech Synthesis Markup Language) messages.
327 * Communication with Festival is highly improved.
328 * Improved DTK Software Synthesis support.
329 * Python interface library
330 * Bugfixes
331
332 Version 0.4
333 * Volume settings in the Generic output module.
334 * The 16-bit voice for Flite is installed if available.
335 * spd-say now accepts the most important speech parameters.
336 * libspeechd now uses TCP_NODELAY to make the communication faster.
337 * Bugfixes.
338
339 Version 0.3
340 * Volume setting commands in SSIP.
341 * Package and binary renamed to speech-dispatcher.
342 * Generic output module configuration file for DTK Software Synthesis.
343 Thanks to Jacob Schmude.
344 * Document with definition of SSIP is now separated.
345 * Debugging.
346
347 Version 0.2
348 * Improved documentation (definition of the SSIP
349 protocol is now a separated document)
350 * Caching of characters, keys and sound icons in the
351 Festival output modules brings much quicker responses
352 for these types of events.
353 * If some output module fails to load completely (e.g. because
354 it's unable to connect to the synthesizer), error is reported
355 to Speech Dispatcher and written into the log file. (Sadly,
356 this doesn't work with kernels that don't support real threading).
357 * Necessary recoding in Festival output module is now retrieved
358 from Festival itself.
359
360 Version 0.1
361 * First official release
362 * Improved documentation
363 * We have abandoned the idea to do some text processing in Speech Dispatcher,
364 it's all handled in the output modules
365 * New scheme scripts so that Festival supports some of the SSIP commands.
366 * Improvements in the C-library
367 * Documentation in Czech.
368 * Debugging, testing.
369
370 Version 0.0.10
371 * Output modules are now external programs communicating through stdin/stdout
372 * Output modules can be reloaded at run time sending SIGUSR1 to speechd
373 * Multiple generic modules loaded at once supported
374 * Filtering in output layer of Speech Dispatcher is now optional
375 and can be handled later in output_modules or in the TTSs they use
376 * Client specific configuration
377 * Bug-fixes
378
379 Version 0.0.9
380 * Generic output module (+definition for the Epos synthesiser)
381 * Better autoconf configuration
382 * Restructured documentation
383 * Block support in SSIP
384 * Daemon implementation
385 * Better configuration
386 * Capital letters recognition
387 * Experimental support for ogg sound icons.
388 * Bug-fixes
389
390 Version 0.0.8
391 * Output modules reworked (now using a common mechanism)
392 * New priority system (now using 5 priorities)
393 * Improved documentation
394 * Improved C API
395 * Tables for czech language
396 * run_test and new tests, sayfortune.sh (src/tests/)
397 * Many bugs fixed
398 * Renamed to Speech Dispatcher
399
400 Version 0.0.7
401 * Support for sound-mapped icons (cstsnd output module)
402 * Debian package support
403 * Apollo output module
404 * Experimental Python API
405 * Tests in src/tests/
406 * Improved PAUSE mechanism
407
408 Version 0.0.6
409 * Speech Dispatcher now works in 2 threads for better performance
410 * New output module for Flite (using Flite C API)
411 * Added debug mode
412 * Spelling, punctuation, CHAR and KEY commands
413 * Heavily improved documentation of SSIP (thx Milan Zamazal)
414 * Chages in SSIP
415 * Emacspeak server
416 * Python API
417
418 Version 0.0.5
419 * Renamed back to Speech Dispatcher
420 * Speech Dispatcher Center -- history and configuration client (first version)
421 * command line / menu logic
422 * browsing through clients
423 * browsing through client's history
424 * stop/pause/resume functions
425 * speech support
426 * command line completion
427 * Lower CPU usage
428 * Sound icons support
429
430 Version 0.0.4
431 * Automake and autoconf support
432 * Robod now uses GLib2 instead of GDSL
433 * More polished source
434 * Improved configuration of server
435 * More functions in the API.
436
437 Version 0.0.3
438 * API for communication with Robod.
439 * New output module for flite.
440 * History commands implemented.
441 * Shared library API in Robodlib.so.
442
443 Version 0.0.2
444
445 * API for output modules designed
446 * sample output module implementation written
447 * server configuration based on DotConf implemented
448 * analysis of input protocol requirements documented
449
450 Version 0.0.1
451
452 * Project started ...
453 * Design documentation written (using DocBook XML)
454 * pre API designed
455 * simple server core and client example written
456 * CVS repository and other project stuff set up
457
458
459 Copyright (C) 2002-2012 Brailcom, o.p.s
460 Copyright (C) 2002-2012 Brailcom, o.p.s
461 Copyright (C) 2018-2025 Samuel Thibault <[email protected]>
462 Copyright (C) 2018 Florian Steinhardt <[email protected]>
463
464 This program is free software; you can redistribute it and/or modify it under
465 the terms of the GNU General Public License as published by the Free Software
466 Foundation; either version 2 of the License, or (at your option) any later
467 version.
468
469 This program is distributed in the hope that it will be useful, but WITHOUT ANY
470 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
471 PARTICULAR PURPOSE. See the GNU General Public License for more details (file
472 COPYING in the root directory).
473
474 You should have received a copy of the GNU General Public License
475 along with this program. If not, see <https://www.gnu.org/licenses/>.