208
208
If this doesn't work, please test again if all the components work
209
209
well as described above.
211
10. It's very slow etc...
211
10. It's somehow slow etc...
213
213
You might find the new software synthesis is rather slow with
214
214
SpeakUp compared to the hardware synthesizers, and they have some
215
other deficiencies. This is not a fault of Speech Dispatcher or
216
software synthesis in general. It might be the fault of your
217
software synthesizer, but most probably (as with Festival and
218
others), it's because of the way SpeakUp does interfacing to
219
software synthesis. SpeakUp was not formerly designed to
220
interface with software synthesis and neither now this is its main
221
focus. Please don't think software synthesis is inferior to
222
hardware synthesis because of this. See *Note Problems::
215
other deficiencies. On today's hardware, this is not a fault of
216
Speech Dispatcher or software synthesis in general. Speakup was
217
not formerly designed with software synthesis in mind and neither
218
now this is its main focus. Please don't think software synthesis
219
is generally inferior to hardware synthesis because of this. We
220
actually believe quite the opposite is true: The hardware
221
synthesizer, unless they are programable, will never achieve the
222
huge possibilities of extensible software synthesizers. See *Note
247
248
Selects the device where Speakup sends it's output.
250
Indicates which character coding Speakup uses. For possible
251
values, please see `iconv -list'
251
Indicates which character coding your console uses. For possible
252
values, please see `iconv -list'. This option is important if your
253
console is not in iso-8859-1! (It doesn't work with "utf-8" and
254
other multibyte encoding because of kernel related issues with this
255
encoding. You must run your console in a unibyte encoding like for
256
example one of the "iso-8859-*" set, sorry.)
258
`-t or --dont-init-tables'
259
SpeechD-Up tries to init some Speakups /proc tables on its start.
260
Most important, it changes /proc/characters and /proc/chartab so
261
that punctuation, capital letters recognition and various other
262
language dependent things are handled inside Speech Dispatcher and
263
in the synthesizer. This is crucial for internationalization. You
264
can however disable this if you want to modify the tables manually
265
for some reason and don't like speechd-up overwriting them. (Do
266
not expect Speakup to work for other languages than English in
267
that case, unless you know what you are doing and can do the
268
necessary changes manually.)
254
271
Runs SpeechD-Up in the probe mode. This means that SpeechD-Up will
255
do everythingas as ordinary, but won't try to open the SpeakUp
272
do everything as as ordinary, but won't try to open the SpeakUp
256
273
device. It just speaks a message and terminates (indicating so in
257
274
the logfiles). This is meant for testing.
267
284
If the device where Speakup talks to userspace is `/dev/sftsyn2' and
268
the input is in `utf-8', then you can use (this doesn't actually work
269
now, see the section *Note Problems::):
285
your console is in `iso-8859-2', then you can use:
271
speechd-up --device="/dev/sftsyn2/" --coding "utf-8"
287
speechd-up --device="/dev/sftsyn2/" --coding "iso-8859-2"
273
289
Please use this for debugging or sending bug-reports
274
290
speechd-up -s -l 5 -L "/tmp/speechd-up.log"
282
There are many known problems with Speakup/SpeechD-Up. You might want to
283
consider using some user-space assistive technologies if you want to use
284
software synthesis. Such technologies are speechd-el, Emacspeak, SuSE
287
* Reading keys/characters is slow
289
The way Speakup passes individual characters after the user types
290
them on the screen makes it impossible to tell these are
291
individual characters typed by the user and so no caching
292
mechanisms are used in Speech Dispatcher and everything is
293
synthesized like an ordinary text.
295
Worse, spaces appearing around single characters as received by
296
Speakup make some synthesizers add additional delays into the
299
* It doesn't speak my language / it uses some weird character set
301
Speakup currently only supports English. You can make a try with
302
speechd-up -c <character-set> but if you don't succeed, it's
303
better to give it up (unless you want to contribute to Speakup's
298
There are many known problems with Speakup/SpeechD-Up. You might want
299
to consider using some user-space assistive technologies if you want to
300
use software synthesis on the console. Such technologies are
301
speechd-el, Emacspeak, SuSE Blinux and Yasr. We however believe the
302
Speakup/SpeechD-Up combination is useful even though it has the
303
described deficiencies.
305
* (Fixed: Speaking characters and keys is slow)
307
The original issue was fixed (but see bellow). If you are still
308
not happy with the performance while reading characters and typing
309
in, then you are using a synthesizer which is both slow and
310
lacking support for caching (use Festival, it does caching with
311
Speech Dispatcher and keys reading is instant).
313
* Characters behave like KEYS for synthesizer configuration
315
The way Speakup passes individual characters makes it impossible
316
to tell whether this character originated as a key or while the
317
user was moving the cursor on the screen and the already present
318
letters are being read. We believe it is important to carry the
319
proper information when pressing keys (like capitalization), so
320
SpeechD-Up handles these messages as keys.
322
Also space, tab and various other whitespace characters are not
323
reported as a characters/keys as this is impossible to achieve
324
with current Speakups implementation. It is impossible if the word
325
"space" in input originated by the user pressing the SPACE key or
326
by encountering this word inside a document being read on the
329
* (Fixed: It doesn't speak my language / it uses some weird
332
If you still experience this issue, you are probably doing one of
335
* Using a wide character encoding like UTF-8. This is not
336
supported for inherent issues with wide characters inside the
339
* Not running `speechd-up' with the `-c' parameter and
340
specifying the proper character set your console uses.
342
* Running `speechd-up' with the `-t' or `--dont-init-tables'.
343
This option is incompatible with internationalization.
345
If this is not true, please contact us on
346
<speechd@bugs.freebsoft.org>.
306
348
* The spelling/punctuation modes of my favorite synthesizer don't
315
357
UTF-8 should work as long as you only use characters encountered
316
358
in the basic ASCII. So English should work. But Speakup doesn't
317
359
really support UTF-8 in any way (nor it can, there is no UTF-8
318
support inside the kernel - just workarounds targeted at visuall
360
support inside the kernel - just workarounds targeted at visual
319
361
interpretation; the kernel virtual terminal drivers contain many
320
362
barriers for accessibility).
322
* I can't switch synthesizers on fly
364
* Punctuation and capital letters recognition doesn't properly work
366
This is an issue again related to Speakups design. It is currently
369
* I can't switch languages or synthesizers on fly
324
371
The way the interface Speakup - Speech Dispatcher is designed,
325
372
this is not possible. You can switch them statically in Speech
328
375
specifically for Speakup.
330
377
* Any other more advanced thing doesn't work or something is too
331
slow or ... There are many problems with how this interface is
332
done and with how Speakup is designed. It's probably much better
333
to use Speakup with a hardware synthesizer, if you have the
334
possibility. Speakup was not designed with user space software
335
synthesis in mind and the necessary changes would probably be too
336
difficult now. There were many important design decisions made
337
which are very reasonable for hardware synthesizers, but they just
338
don't mix well with software synthesis.
378
slow or ... There are many problems for software speech with how
379
this interface is done and with how Speakup is designed. It's
380
probably still better to use Speakup with a hardware synthesizer,
381
if you have the possibility. Speakup was not designed with user
382
space software synthesis in mind and the necessary changes would
383
probably be too difficult now. There were many important design
384
decisions made which are very reasonable for hardware
385
synthesizers, but they just don't mix well with software synthesis.
340
387
On the other hand, experience shows people still find this
341
388
interface to software speech synthesis useful, mostly because of
342
lack of anything else. That's the reason why we provide this
389
lack of anything else. Also, it seems the future of accessibility
390
technologies lies in the software synthesis (extensible, etc.) as
391
machines get faster. That's the reason why we provide this
343
392
application and we will try to help you if you have problems or
348
396
File: speechd-up.info, Node: Contact and Reporting Bugs, Next: How You Can Help, Prev: Problems, Up: Top
393
441
* _Help us with project:_ Please look at `http://www.freebsoft.org/'
394
442
to find information about our projects. There is a plenty of work
395
to be done for the blind and visually impaired people to make
396
their work with computers easier.
443
still to be done for the blind and visually impaired people to
444
make their work with computers easier.
398
446
* _Spread a word about Speech Dispatcher and Free Software:_ You can
399
447
help us, and the whole community around Free Software, just by