I got Espeakup finally to work
Hi, just let you know: Espeakup now works in German. I had to pull the latest from GIT, and after compiling I had to type: espeakup -V de or espeakup --default-voice=de Note: You can replace that "de" by any non-English voice. I got the tip from the Speakup list archive. Question: Is it possible to replace the 0.4 by the latest version (0.60)? Second note: the modified characters file is still needed. Hermann
--
Good to hear you've got it working. While you didn't ask for it, I can imagine it would be useful, would you like the swspeak script to identify your language and start espeakup with the correct -V option? My idea is that all this stuff should be as transparent to the user as possible (IE. you've set your language, the console may be using your language, so why should you need to set extra for your speech to speak your language). Or would it be better to be an espeakup enhancement to auto detect your language? If I get time I may look at this.
Any thoughts, personally I would have thought espeakup would be the better place although swspeak script may be easier for me to modify.
Michael Whapples On -10/01/37 20:59, Hermann wrote:
Hi, just let you know: Espeakup now works in German. I had to pull the latest from GIT, and after compiling I had to type: espeakup -V de or espeakup --default-voice=de Note: You can replace that "de" by any non-English voice. I got the tip from the Speakup list archive. Question: Is it possible to replace the 0.4 by the latest version (0.60)? Second note: the modified characters file is still needed. Hermann
On 28.05.2009 at 12:28:13 Michael Whapples mwhapples@aim.com wrote:
Good to hear you've got it working. While you didn't ask for it, I can imagine it would be useful, would you like the swspeak script to identify your language and start espeakup with the correct -V option? My idea is that all this stuff should be as transparent to the user as possible (IE. you've set your language, the console may be using your language, so why should you need to set extra for your speech to speak your language). Or would it be better to be an espeakup enhancement to auto detect your language? If I get time I may look at this.
I guess it would be better to implement an autodetection of the locale in Espeakup, since swspeak not only starts Espeakup, but also Speechd-up, where, maybe, a different approach is needed. Note that Speechd-up almost detects the locale correctly, and only the characters file must be edited; Espeakup failed completely till 0.6 and the characters file has to be edited as well.
Any thoughts, personally I would have thought espeakup would be the better place although swspeak script may be easier for me to modify.
See above. But what about that characters file? I guess this is due to Speakup. It should be possible that Speakup identifies the locale on start up. Has this been discussed on the Speakup list? Hermann
I don't know whether locale detection in speakup has been discussed on the speakup list, I don't know whether speakup can do locale detection or not as it is kernel modules. If I don't see a message from you on the speakup list soon I will ask for you and try and find out.
Regarding the replacing of speakup's tables, I thought speechd-up did that, may be I am thinking of 0.4, have you tried updating speechd-up to that (it would require compiling it from source as GRML comes with speechd-up 0.3). My thought is that may be the thinking is for synths/synth connectors to set up the speakup table to work with them (eg. by default speakup's characters table says "zehd" for the letter z,this is due to Kirk the author of speakup being Canadian and like us in the UK don't like the American "zee" of US synths, however due to this espeak says the letter z unusually when using speakup, this indicates the table may need to be synth specific).
Michael Whapples On 28/05/09 11:50, Hermann wrote:
On 28.05.2009 at 12:28:13 Michael Whapplesmwhapples@aim.com wrote:
Good to hear you've got it working. While you didn't ask for it, I can imagine it would be useful, would you like the swspeak script to identify your language and start espeakup with the correct -V option? My idea is that all this stuff should be as transparent to the user as possible (IE. you've set your language, the console may be using your language, so why should you need to set extra for your speech to speak your language). Or would it be better to be an espeakup enhancement to auto detect your language? If I get time I may look at this.
I guess it would be better to implement an autodetection of the locale in Espeakup, since swspeak not only starts Espeakup, but also Speechd-up, where, maybe, a different approach is needed. Note that Speechd-up almost detects the locale correctly, and only the characters file must be edited; Espeakup failed completely till 0.6 and the characters file has to be edited as well.
Any thoughts, personally I would have thought espeakup would be the better place although swspeak script may be easier for me to modify.
See above. But what about that characters file? I guess this is due to Speakup. It should be possible that Speakup identifies the locale on start up. Has this been discussed on the Speakup list? Hermann
On 28.05.2009 at 13:38:56 Michael Whapples mwhapples@aim.com wrote: [...]
Regarding the replacing of speakup's tables, I thought speechd-up did that, may be I am thinking of 0.4, have you tried updating speechd-up to that (it would require compiling it from source as GRML comes with speechd-up 0.3).
I tried both versions, but the same result: In general, when Speechd-up is used, the German text is spoken well, except the punctuation chars and the numbers, and with the two letters "a" and "z", the latter you mentioned. When moving the cursor char by char, those strange char naming comes up, due to the fact, that chars of >127 are not recognized. The difference to Espeakup is, that, out of the box, it doesn't work at all in German, and when the characters file is adjusted, it spells right when moving cursor-left-right, but it doesn't read the umlauts when reading German text. This all is fixed for Espeak since 0.6 together with the modified characters file. So my conclusion is, that the main issue is in Speakup, since everything works well when I use built-in speech support of Brltty or Suse-Blinux. I hope I could clarify things a bit. Is it possible to write to the Speakup list when one is not subscribed? Because I didn't use Speakup for a while, I left the list. Hermann
Before saying more on this it may be worth reading the speechd-up documentation http://cvs.freebsoft.org/doc/speechd-up/. The problems page may describe what's going on better. I also will have a good read of it before saying more.
I don't know about the possibility of posting to the speakup list without being subscribed, I guess probably not as it has some stupid limitations.
Unless there is anything to do in GRML's swspeak I would suggest possibly moving this to the speakup list.
Michael Whapples On 28/05/09 13:02, Hermann wrote:
On 28.05.2009 at 13:38:56 Michael Whapplesmwhapples@aim.com wrote: [...]
Regarding the replacing of speakup's tables, I thought speechd-up did that, may be I am thinking of 0.4, have you tried updating speechd-up to that (it would require compiling it from source as GRML comes with speechd-up 0.3).
I tried both versions, but the same result: In general, when Speechd-up is used, the German text is spoken well, except the punctuation chars and the numbers, and with the two letters "a" and "z", the latter you mentioned. When moving the cursor char by char, those strange char naming comes up, due to the fact, that chars of>127 are not recognized. The difference to Espeakup is, that, out of the box, it doesn't work at all in German, and when the characters file is adjusted, it spells right when moving cursor-left-right, but it doesn't read the umlauts when reading German text. This all is fixed for Espeak since 0.6 together with the modified characters file. So my conclusion is, that the main issue is in Speakup, since everything works well when I use built-in speech support of Brltty or Suse-Blinux. I hope I could clarify things a bit. Is it possible to write to the Speakup list when one is not subscribed? Because I didn't use Speakup for a while, I left the list. Hermann
* Michael Whapples mwhapples@aim.com [20090528 12:28]:
Good to hear you've got it working. While you didn't ask for it, I can imagine it would be useful, would you like the swspeak script to identify your language and start espeakup with the correct -V option? My idea is that all this stuff should be as transparent to the user as possible (IE. you've set your language, the console may be using your language, so why should you need to set extra for your speech to speak your language). Or would it be better to be an espeakup enhancement to auto detect your language? If I get time I may look at this.
Any thoughts, personally I would have thought espeakup would be the better place although swspeak script may be easier for me to modify.
grml has all the relevant options for handling language stuff, so if espeakup just needs to be started with some special options when using german settings (using lang=de) with swspeak this shouldn't be a big deal, yes.
Michael, you seem to be very good at all the swspeak/espeakup/... related stuff. Are you interested in helping us in better integration of swspeak/espeakup/... at grml? I don't have any active developers working on accessibility in grml (so far I'm doing all that stuff on my own), so I'd highly welcome you as a contributor.
regards, -mika-
Before saying yes, I will just state I need to learn about debian packaging. The only debian packages I have created have been on my own python packages using stdeb. Any guides you would suggest for debian packaging?
Also I will briefly mention a little about myself.
I do use speakup (I have an apollo synth on my desktop and for my laptop I use software speech output with espeakup). I also have a Braille display which I use brltty with.
My view on accessibility is that where ever possible it should be in the main distro and easy to get running. When I say easy to get running I realise it can't be fully automatic on a CD such as GRML as not all users want it, but ideally giving the swspeak command at the boot prompt should be enough for software speech to come up working for the user with all language settings as they gave (IE. if the console shows german then speech should be german). On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me. Related to this, when I do grml2hd the HD installation when it starts will not come up automatically speaking, even if I used swspeak for install. This has caught me and I think others out in the past, as the instructions from the speech messages tell you type swspeak once booting finishes but the HD installation puts you at a login screen so you need to log in first.
So I am saying yes at the moment, but I would like to know fully what is involved before saying yes for definite.
Michael Whapples
On -10/01/37 20:59, Michael Prokop wrote:
- Michael Whapplesmwhapples@aim.com [20090528 12:28]:
Good to hear you've got it working. While you didn't ask for it, I can imagine it would be useful, would you like the swspeak script to identify your language and start espeakup with the correct -V option? My idea is that all this stuff should be as transparent to the user as possible (IE. you've set your language, the console may be using your language, so why should you need to set extra for your speech to speak your language). Or would it be better to be an espeakup enhancement to auto detect your language? If I get time I may look at this.
Any thoughts, personally I would have thought espeakup would be the better place although swspeak script may be easier for me to modify.
grml has all the relevant options for handling language stuff, so if espeakup just needs to be started with some special options when using german settings (using lang=de) with swspeak this shouldn't be a big deal, yes.
Michael, you seem to be very good at all the swspeak/espeakup/... related stuff. Are you interested in helping us in better integration of swspeak/espeakup/... at grml? I don't have any active developers working on accessibility in grml (so far I'm doing all that stuff on my own), so I'd highly welcome you as a contributor.
regards, -mika-
On 02.06.2009 at 15:23:44 Michael Whapples mwhapples@aim.com wrote: [...]
On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me.
And the question is also, why does that second "swspeak" open a new shell? It should stay to the shell opened at startup.
Related to this, when I do grml2hd the HD installation when it starts will not come up automatically speaking, even if I used swspeak for install. This has caught me and I think others out in the past, as the instructions from the speech messages tell you type swspeak once booting finishes but the HD installation puts you at a login screen so you need to log in first.
This is new to me, but I never started the grml2hd with speech, I used braille and this does not happen. Perhaps because I started the script using some parameters, such as the target drive and the place where the boot manager has to be written. Hermann
On 02/06/09 14:58, Hermann wrote:
On 02.06.2009 at 15:23:44 Michael Whapplesmwhapples@aim.com wrote: [...]
On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me.
And the question is also, why does that second "swspeak" open a new shell? It should stay to the shell opened at startup.
So that's what happens, I had noticed the prompt is different between finished booting and after running swspeak.
Related to this, when I do grml2hd the HD installation when it starts will not come up automatically speaking, even if I used swspeak for install. This has caught me and I think others out in the past, as the instructions from the speech messages tell you type swspeak once booting finishes but the HD installation puts you at a login screen so you need to log in first.
This is new to me, but I never started the grml2hd with speech, I used braille and this does not happen. Perhaps because I started the script using some parameters, such as the target drive and the place where the boot manager has to be written.
May be I was unclear about what I meant there. When I install GRML to my HD it picks up about the swspeak boot option and adds it to lilo. The problem is that swspeak as a boot option still requires you to type swspeak once booting finishes. Now on the Live CD you don't get presented with a log in screen therefore you can just start typing swspeak, but on HD it requires you to log in first and then you can run swspeak command. Its that the speech messages are the same as those on the Live CD and so don't quite match the HD installation. Braille and hardware speech support aren't affected as those options don't require anything to be entered once booting finishes for them to work (IE. brltty is started when you give the brltty or blind option at the boot prompt). Its this having to run a command once booting finishes which causes the problem.
Michael Whapples
Hermann
* Hermann meinelisten@onlinehome.de [20090602 15:56]:
On 02.06.2009 at 15:23:44 Michael Whapples mwhapples@aim.com wrote: [...]
On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me.
And the question is also, why does that second "swspeak" open a new shell? It should stay to the shell opened at startup.
Well, that's not really true for the current grml version anymore, it's just that special options for Zsh are set up, because the default ones are known to cause problems for some of you.
regards, -mika-
* Michael Whapples mwhapples@aim.com [20090602 15:24]:
Before saying yes, I will just state I need to learn about debian packaging. The only debian packages I have created have been on my own python packages using stdeb. Any guides you would suggest for debian packaging?
It's not that important that you do the packaging for us. It would help me a lot if you'd (kind of regularly) test our stuff, help us in accessibility related questions and provide suggestions and tips how we could improve stuff. It's especially hard for me to check out all the speakup related features because I don't have the hardware on my own and don't really know what people using those stuff really need.
Regarding packaging: have a look at the official Debian devel documentation:
Debian New Maintainers' Guide: http://www.debian.org/doc/maint-guide/
Debian Developer's Reference: http://www.debian.org/doc/developers-reference/
Debian Policy Manual: http://www.debian.org/doc/debian-policy/
Also I will briefly mention a little about myself.
I do use speakup (I have an apollo synth on my desktop and for my laptop I use software speech output with espeakup). I also have a Braille display which I use brltty with.
Ok fine.
My view on accessibility is that where ever possible it should be in the main distro and easy to get running. When I say easy to get running I realise it can't be fully automatic on a CD such as GRML as not all users want it, but ideally giving the swspeak command at the boot prompt should be enough for software speech to come up working for the user with all language settings as they gave (IE. if the console shows german then speech should be german). On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me. Related to this, when I do grml2hd the HD installation when it starts will not come up automatically speaking, even if I used swspeak for install. This has caught me and I think others out in the past, as the instructions from the speech messages tell you type swspeak once booting finishes but the HD installation puts you at a login screen so you need to log in first.
Alright, regarding the swspeak setup (the "type swspeak once booting finished") I've created a wishlist bugreport in our bug tracking system: http://bts.grml.org/grml/issue684
Regarding grml2hd: I'm *not* a fan of the fact that people that don't know how to deal with Debian/unstable install a Debian/unstable based system on their box. There are several Debian developers working in the accessibility team of Debian and I'd like to have good accessibility in the official Debian installer and distribution. grml should be a good environment for all of you as a *live* system but if you intend to install Debian with accessibility features to harddisk then please either use the official Debian installer or consider using grml-debootstrap (which installs plain Debian on your system). But please do *not* use grml2hd if you're not really familiar with Debian/unstable.
So what I'd like to know: what are the biggest problems for you in using the official Debian installer with regards to accessibility? Might it be worth the effort to put work into grml-debootstrap providing an official Debian accessibility enabled harddisk installation?
And I'd like to work out a list of what software is important for you nowadays. speakup, speech-dispatcher, espeakup, speechd-up,... etc. - I somehow lost the overview what's the way to go nowadays, what needs an update, what could be dropped, etc.
regards, -mika-
On -10/01/37 20:59, Michael Prokop wrote:
- Michael Whapplesmwhapples@aim.com [20090602 15:24]:
[...]
It's not that important that you do the packaging for us. It would help me a lot if you'd (kind of regularly) test our stuff, help us in accessibility related questions and provide suggestions and tips how we could improve stuff. It's especially hard for me to check out all the speakup related features because I don't have the hardware on my own and don't really know what people using those stuff really need.
Fine with me. I just thought that knowledge on packaging would be useful even if so I could say why the package is wrong. Will have a read of the docs you refer to below.
Regarding packaging: have a look at the official Debian devel documentation:
Debian New Maintainers' Guide: http://www.debian.org/doc/maint-guide/
Debian Developer's Reference: http://www.debian.org/doc/developers-reference/
Debian Policy Manual: http://www.debian.org/doc/debian-policy/
[...]
My view on accessibility is that where ever possible it should be in the main distro and easy to get running. When I say easy to get running I realise it can't be fully automatic on a CD such as GRML as not all users want it, but ideally giving the swspeak command at the boot prompt should be enough for software speech to come up working for the user with all language settings as they gave (IE. if the console shows german then speech should be german). On this topic it is something I never really worked out, why does GRML require me to type swspeak once booting finishes as surely the start up script which gives the speech messages (eg. "software speech has been enabled, type swspeak once booting finishes" or something like that) could run the swspeak command for me. Related to this, when I do grml2hd the HD installation when it starts will not come up automatically speaking, even if I used swspeak for install. This has caught me and I think others out in the past, as the instructions from the speech messages tell you type swspeak once booting finishes but the HD installation puts you at a login screen so you need to log in first.
Alright, regarding the swspeak setup (the "type swspeak once booting finished") I've created a wishlist bugreport in our bug tracking system: http://bts.grml.org/grml/issue684
Will look at that report and put further comments on that as necessary.
Regarding grml2hd: I'm *not* a fan of the fact that people that don't know how to deal with Debian/unstable install a Debian/unstable based system on their box. There are several Debian developers working in the accessibility team of Debian and I'd like to have good accessibility in the official Debian installer and distribution. grml should be a good environment for all of you as a *live* system but if you intend to install Debian with accessibility features to harddisk then please either use the official Debian installer or consider using grml-debootstrap (which installs plain Debian on your system). But please do *not* use grml2hd if you're not really familiar with Debian/unstable.
OK, my grml2hd comment was partly that messages currently are wrong once performed the installation due to the difference of having to login to the installed system but no need to login on a LiveCD. If software speech is all automatically started then this will not cause a problem as the speech output from speakup will let users know that they are at a login screen.
Relating to your comments about grml2hd not being recommended for those unfamiliar with debian unstable. I plan to make an audio walkthrough for using GRML to install linux with software speech, can I infer that you would suggest this walk through should be using grml-debootstrap, as users who grml2hd would be suitable for would be fine with reading docs and working out what to do.
So what I'd like to know: what are the biggest problems for you in using the official Debian installer with regards to accessibility? Might it be worth the effort to put work into grml-debootstrap providing an official Debian accessibility enabled harddisk installation?
In the past there hasn't been an official debian installation CD with software speech output. I think there was a installer CD back around kernel 2.6.18 but this only used speakup with hardware speech output. I think there may be some work being done on creating a suitable debian installer CD with software speech output, but I have stuck with GRML as it is also such a good live environment which provides me with all the accessibility I need.
And I'd like to work out a list of what software is important for you nowadays. speakup, speech-dispatcher, espeakup, speechd-up,... etc. - I somehow lost the overview what's the way to go nowadays, what needs an update, what could be dropped, etc.
I personally feel espeakup gives me a better user experience than speech-dispatcher and speechd-up. Reasons are: espeak and espeakup seem to be very responsive and tend not to split words up (speech-dispatcher on the GRML 1.0 or something like that suffered from problems like that), speech-dispatcher is a server which can and sometimes does die on me so leaving me without speech (espeakup as far as I can remember has never died on me) and espeakup is certainly getting frequent updates (speechd-up seems to be slower on development). Speech-dispatcher has the following advantages: I think from what people say seems to be quite good for internationalisation (I am in the UK so use english, so don't find some of the internationalisation issues), speech-dispatcher seems to be being pushed for becoming a speech API for linux (I think there are plans for gnome-speech to be dropped for gnome 3.0 and speech-dispatcher to take its place) and other synthesisers can be used by speech-dispatcher (espeakup is tied to espeak). Its a shame the speech side of things isn't as clear as it is with the braille display support, brltty.
As a note about espeakup, the one used by GRML is 0.41 (or something around that version, quite old) but debian seem to have an espeakup package now for 0.71 (although package is numbered relative to speakup, 3.0..4.dfsg.1-2). As debian is now providing this package and the debian one is newer may be follow that.
Michael Whapples
regards, -mika-
Teilnehmer (3)
-
Hermann -
Michael Prokop -
Michael Whapples