Hi, Is there a program that can convert text to voice? Even better, use the PC loudspeaker, instead of the soundcard? I think I remember some one mentioning a speaking clock :-? -- Cheers, Carlos Robinson
Carlos E. R. wrote:
Is there a program that can convert text to voice?
Even better, use the PC loudspeaker, instead of the soundcard?
I think I remember some one mentioning a speaking clock :-?
Hi Carlos Does this need to be something for which there is a SuSE RPM or are you happy about doing a bit of compiling yourself? Since it could be done with a Sinclair ZX Spectrum, I guess that it is theoretically possible to create speech through the system speaker. It may well have been done. However, all of the TTS (Text to Speech) systems of which I am aware utilise the system sound server (Alsa, etc.) I am developing an accessible kiosk <http://www.tivis.net.au> which uses the Festival system for text-to-speech work. Please feel free to contact me off-list if you would like to discuss your application - I will see what I can do to help. Ciao M -- Matthew Smith Kadina Business Consultancy South Australia http://www.kbc.net.au
I am developing an accessible kiosk <http://www.tivis.net.au> which uses the Festival system for text-to-speech work.
Hello, I know this has nothing to do with any of your messages, but I am using Suse 8.2 and cannot compile the speech tools portion of festival. Make complains that it cannot find strstream.h and I have all the c and c++(gcc-c++-3.3.1-16, libgcc-3.3.1-16, gcc-3.3.1-16, cpp-3.3.1-16, libstdc++-3.3.1-16, libstdc++-devel-3.3.1-16) development files installed. I was just wondering, if someone could point me into the right direction to get this installed. Thanks and sorry to interrupt.
On Friday 02 January 2004 00:11, Bryan Brown wrote:
Hello, I know this has nothing to do with any of your messages, but I am using Suse 8.2 and cannot compile the speech tools portion of festival. Make complains that it cannot find strstream.h and I have all the c and c++(gcc-c++-3.3.1-16, libgcc-3.3.1-16, gcc-3.3.1-16, cpp-3.3.1-16, libstdc++-3.3.1-16, libstdc++-devel-3.3.1-16) development files installed. I was just wondering, if someone could point me into the right direction to get this installed. Thanks and sorry to interrupt.
I got the same problem, and the point is that the festival project hasn't been updated in a year. The last date I saw, was 2002 and unfortunately too much of the C++ codebase is changing too fast. Just the above case, has changed from 'String' to 'stringstream' to 'strstream' and should now be 'sstream'. These changes don't make things easy, and there are a lot of programs out there that aren't in line with the new template library conventions. But you can try using 'sstream' instead in the source code.
The Friday 2004-01-02 at 08:36 +1030, Matthew Smith wrote:
Carlos E. R. wrote:
Is there a program that can convert text to voice?
Even better, use the PC loudspeaker, instead of the soundcard?
I think I remember some one mentioning a speaking clock :-?
Hi Carlos
Does this need to be something for which there is a SuSE RPM or are you happy about doing a bit of compiling yourself?
Doesn't matter. An rpm is faster, but it doesn't matter if I have to compile something - I do that quite often.
Since it could be done with a Sinclair ZX Spectrum, I guess that it is theoretically possible to create speech through the system speaker.
Some old games did it on Dos, so it is possible - even with 8 Mhz cpus. No, I must be mistaken: I think it was recorded voice. Some one even played music using a pin printer!
It may well have been done. However, all of the TTS (Text to Speech) systems of which I am aware utilise the system sound server (Alsa, etc.)
It's far easier. But perhaps using recorded sounds things like a talking clock can be done using the speaker. Somewhere I read about a sound driver for Linux using the internal speaker, but I never could try it. I could use it for my system warnings, instead of using the soundcard that might be busy, or the speakers could be powered off.
I am developing an accessible kiosk <http://www.tivis.net.au> which uses the Festival system for text-to-speech work.
I'll have a look. :-)
Please feel free to contact me off-list if you would like to discuss your application - I will see what I can do to help.
Nothing in particular - I'm just thinking of simple things like playing some messages from a cron job, or from a script, to remind to do something. Nothing serious. I suddenly remembered that when I bought my first sound card some years ago (a sound blaster) it had a program that could read ascii files and read them aloud (and I think not only in English, but I'm not sure) so I wondered if there was anything similar in Linux. What is festival? I can't find any docs or file about it in SuSE - and I'm offline right now, can't google it ;-) -- Cheers, Carlos Robinson
Carlos E. R. wrote:
Is there a program that can convert text to voice?
I am developing an accessible kiosk <http://www.tivis.net.au> which uses the Festival system for text-to-speech work.
What is festival? I can't find any docs or file about it in SuSE - and I'm offline right now, can't google it ;-)
Festival comes from the University of Edinburgh (where they have the Edinburgh Festival ;-) and is based on the Edinburgh Speech Tools Library. You can find details at <http://www.cstr.ed.ac.uk/projects/speech_tools/> I have been doing my tests using Perl and XSLT to convert well-formed XHTML Web pages into the Sable speech markup language that Festival understands. Although I haven't spent much time on it (I'm hoping to get funding), my initial trials were very promising, including a small Perl programme that could ftp the weather forecast from the Australian Bureau of Meteorology and then read it to me... Throughout my development process, I've run Festival under Suse 8.1 and 8.2 with no problems. A couple of people have reported problems compiling the Speech Tools part, so I'm just compiling it all again on 9.0; no errors compiling Speech Tools, but when I come to compile Festival, the same error shows. I think, however, I have found the source of the error and have fixed it by renaming a header file and leaving a blank on in it's place. Cheers M -- Matthew Smith Kadina Business Consultancy South Australia http://www.kbc.net.au
The Friday 2004-01-02 at 15:03 +1030, Matthew Smith wrote:
What is festival? I can't find any docs or file about it in SuSE - and I'm offline right now, can't google it ;-)
Festival comes from the University of Edinburgh (where they have the Edinburgh Festival ;-) and is based on the Edinburgh Speech Tools Library. You can find details at <http://www.cstr.ed.ac.uk/projects/speech_tools/>
I found it out soon after asking, and I downloaded it - I wonder why SuSE doesn't include it? - I still haven't tried to compile it; version 1.4.3: January 2003, acording to the web page: aREADME-1-4.3-RELEASE festival-1.4.3-release.tar.gz festlex_CMU.tar.gz festlex_POSLEX.tar.gz festvox_ellpc11k.tar.gz festvox_kallpc16k.tar.gz speech_tools-1.2.3-release.tar.gz
I have been doing my tests using Perl and XSLT to convert well-formed XHTML Web pages into the Sable speech markup language that Festival understands.
Although I haven't spent much time on it (I'm hoping to get funding), my initial trials were very promising, including a small Perl programme that could ftp the weather forecast from the Australian Bureau of Meteorology and then read it to me...
Ah, sounds interesting :-)
Throughout my development process, I've run Festival under Suse 8.1 and 8.2 with no problems. A couple of people have reported problems compiling the Speech Tools part, so I'm just compiling it all again on 9.0; no errors compiling Speech Tools, but when I come to compile Festival, the same error shows. I think, however, I have found the source of the error and have fixed it by renaming a header file and leaving a blank on in it's place.
Could you be more specific? It would save me trouble :-) -- Cheers, Carlos Robinson
participants (4)
-
Bryan Brown -
Carlos E. R. -
Matthew Smith -
Örn Hansen