Possible work around to the hardware synthesizer problem

Discussion:

Shawn Kirkpatrick

2016-02-26 01:17:09 UTC

I've written a program that will allow hardware synthesizers to be used with
speakup even thoe the serial support seems to be currently broken. I wrote
this program about a year ago when I thought this problem would be only
temporary. Since it seems like the hardware synthesizer support is still
broken and isn't going to be fixed anytime soon I thought I'd put it out
there in case it can be of some use.
The program is called speakupbridge.
speakupbridge is a program which makes it possible for speakup to use
external serial, parallel, or usb synthesizers. It does this by reading
speakup's softsynth device and passing the text to the synthesizer.
speakupbridge has the following features:
* The ability to communicate with any device that can accept a string
of text using a /dev interface.
* The ability to define the commands used by the synthesizer in a
user-editable configuration file.
* Multiple synthesizer definitions in a single configuration file.
* Change the pronunciation of words using a dictionary file (a feature
speakup
really should do itself).
* Save and reload speakup settings for each defined synthesizer.
For more information or to download the program please visit:
http://www.shawnk.ca/speakup
I haven't had a lot of time to work on or test this code lately so there's
likely to be some rough spots. You'll have to compile the code but that
should be easy enough. I've tested this with my serial Artic transport
synthesizer and it seems to work. I don't use speakup regularly thoe (too
many other missing/bbroken features) so this program really hasn't had any
hard testing.
This solution isn't perfect, you still won't get kernel messages from boot
up but it least it should be possible to use a hardware synthesizer once the
system is started and that's probably better than nothing at all.

c***@ccs.covici.com

2016-02-27 00:26:10 UTC

Permalink

Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

Post by Shawn Kirkpatrick
I've written a program that will allow hardware synthesizers to be
used with speakup even thoe the serial support seems to be currently
broken. I wrote this program about a year ago when I thought this
problem would be only temporary. Since it seems like the hardware
synthesizer support is still broken and isn't going to be fixed
anytime soon I thought I'd put it out there in case it can be of some
use.
The program is called speakupbridge.
speakupbridge is a program which makes it possible for speakup to use
external serial, parallel, or usb synthesizers. It does this by reading
speakup's softsynth device and passing the text to the synthesizer.
* The ability to communicate with any device that can accept a string
of text using a /dev interface.
* The ability to define the commands used by the synthesizer in a
user-editable configuration file.
* Multiple synthesizer definitions in a single configuration file.
* Change the pronunciation of words using a dictionary file (a feature
speakup
really should do itself).
* Save and reload speakup settings for each defined synthesizer.
http://www.shawnk.ca/speakup
I haven't had a lot of time to work on or test this code lately so
there's likely to be some rough spots. You'll have to compile the code
but that should be easy enough. I've tested this with my serial Artic
transport synthesizer and it seems to work. I don't use speakup
regularly thoe (too many other missing/bbroken features) so this
program really hasn't had any hard testing.
This solution isn't perfect, you still won't get kernel messages from
boot up but it least it should be possible to use a hardware
synthesizer once the system is started and that's probably better than
nothing at all.
_______________________________________________
Speakup mailing list
http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici
***@ccs.covici.com

Shawn Kirkpatrick

2016-02-29 08:42:14 UTC

Permalink

I thought of writing a speechdispatcher driver but there were some problems.
When I've tried speechdispatcher with software speech there were lags and
little glitches. I'm not sure if these were being introduced by
speechdispatcher, speechd-up, or some combination of the two. Also, my
version of speechdispatcher has a nasty habbit of segfaulting, not sure why.
I don't think this would solve the indexing problem anyway, as far as I
know speechd-up uses speakup's software synth driver and that doesn't
support indexing, or has this changed? I also think the less layers you have
between speakup and the synth the better, one program is probably better
than two.
What I'd really like to do, if I ever have the time, is write a speach
daemon to replace this whole mess. Something like speechdispatcher but with
more modularity. There could be modules for output, allowing hardware and
software synth support. Modules for input, for various forms of input like
speechdispatcher compatibility, speakup, fifo, or anything else that might
be needed. Modules for conversion, allowing things like a word dictionary,
number processing, etc. The main goal of the program would be to get fast,
responsive speech from whatever synth the user chooses to use.
I think this would be a worthwhile project it would just require time to
write.

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
How do
you spend it?
John Covici

c***@ccs.covici.com

2016-02-29 11:46:19 UTC

Permalink

hmmm, I am using 0.8.3 of speech dispatcher and not seeing any
segfaults or any such and indexing works fine -- although my speechd-up
was compiled from a source which maybe you don't have, I am not sure
where this came from. I was thinking that the output module for a
hardware synth would not be too hard to write and that is why I
suggested it. Your ideais nice, but remember we would want this to work
with orca as well as a client, otherwise for those who use both speakup
and orca, it would be a mess.

I can send you my source for speechd-up, if you would like.

Post by Shawn Kirkpatrick
I thought of writing a speechdispatcher driver but there were some
problems. When I've tried speechdispatcher with software speech there
were lags and little glitches. I'm not sure if these were being
introduced by speechdispatcher, speechd-up, or some combination of the
two. Also, my version of speechdispatcher has a nasty habbit of
segfaulting, not sure why.
I don't think this would solve the indexing problem anyway, as far as I
know speechd-up uses speakup's software synth driver and that doesn't
support indexing, or has this changed? I also think the less layers
you have between speakup and the synth the better, one program is
probably better than two.
What I'd really like to do, if I ever have the time, is write a speach
daemon to replace this whole mess. Something like speechdispatcher but
with more modularity. There could be modules for output, allowing
hardware and software synth support. Modules for input, for various
forms of input like speechdispatcher compatibility, speakup, fifo, or
anything else that might be needed. Modules for conversion, allowing
things like a word dictionary, number processing, etc. The main goal
of the program would be to get fast, responsive speech from whatever
synth the user chooses to use.
I think this would be a worthwhile project it would just require time
to write.

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
How do
you spend it?
John Covici

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici
***@ccs.covici.com

Shawn Kirkpatrick

2016-02-29 14:21:54 UTC

Permalink

I assume you're talking about indexing working with speakup and not orca?
What version of speakup are you using? I don't see how indexing could be
working since the soft synth driver doesn't seem to send any codes that
would allow it to work. Unless things have changed pretty recently. In that
case writing indexing support in to my program probably wouldn't be that
hard. Of course not all hardware synths will do indexing anyway.

Post by c***@ccs.covici.com
hmmm, I am using 0.8.3 of speech dispatcher and not seeing any
segfaults or any such and indexing works fine -- although my speechd-up
was compiled from a source which maybe you don't have, I am not sure
where this came from. I was thinking that the output module for a
hardware synth would not be too hard to write and that is why I
suggested it. Your ideais nice, but remember we would want this to work
with orca as well as a client, otherwise for those who use both speakup
and orca, it would be a mess.
I can send you my source for speechd-up, if you would like.

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
How do
you spend it?
John Covici

c***@ccs.covici.com

2016-02-29 15:24:01 UTC

Permalink

Well, the combination of speakup soft synth, speechd-up, speech
dispatcher and the espeakup driver does work, as far as using speakup-r
goes. I am not sure if its really using indexing, but it has that
effect.

Post by Shawn Kirkpatrick
I assume you're talking about indexing working with speakup and not
orca? What version of speakup are you using? I don't see how indexing
could be working since the soft synth driver doesn't seem to send any
codes that would allow it to work. Unless things have changed pretty
recently. In that case writing indexing support in to my program
probably wouldn't be that hard. Of course not all hardware synths will
do indexing anyway.

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
How do
you spend it?
John Covici

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici
***@ccs.covici.com

c***@ccs.covici.com

2016-02-29 15:25:17 UTC

Permalink

Also, orca's read to end or whatever they call it works as well, just to
let you know.

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
How do
you spend it?
John Covici

--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici
***@ccs.covici.com

John G Heim

2016-02-29 16:20:08 UTC

Permalink

WWhat is indexing?

Post by c***@ccs.covici.com
Shawn, maybe it might be easier and more universal to write a speech
dispatcher driver instead? That way, if you use speechd-up, indexing
would work. What do you think?

--
--
John G. Heim; ***@math.wisc.edu; sip://***@sip.linphone.org

Gregory Nowak

2016-02-29 21:59:29 UTC

Permalink

Indexing is where the speech cursor follows what is being spoken by
the synthesizer. So, if the last thing spoken was "indexing" and you
interrupted speech, the speech cursor would be just after the g of
indexing, or very close to it if both the synthesizer and screen
reader support indexing.

Greg

Post by John G Heim
WWhat is indexing?

--
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)
If we haven't been in touch before, e-mail me before adding me to your contacts.

--
Free domains: http://www.eu.org/ or mail dns-***@EU.org