Help needed in getting text to speech working

I’m trying to get text to speech working and have got stuck. I’ve installed kttsd which also pulled in espeak.

According to the I started the ktts manager which opens and asks me to configure the talkers. The problem is that the list of talkers is empty. I don’t have the items listed in box 2.

Can anyone offer any suggestions?
Also, does phonon need to be installed and enabled for espeak to work. It wasn’t a dependency when it installed but I’ve seen some conflicting references that phonon needs to be installed.

Didn’t pull in espeak for me! Even with it installed the only way I can make it run is by piping into aplay any way.

espeak "hello" --stdout | aplay

Trying festival it works, and I never get espeak as a choice. I wonder whether something lower down is no longer enabled/available. I noticed that kttsmgr had the festival choices and freeTTS even though neither where available on initial testing, looks like its hard coded to me.

I just installed festival and that doesn’t show up in kttsmgr either. I tried un-installing espeak and it wants to downgrade kttsd and remove some stuff.

#### YaST2 conflicts list - generated 2010-02-25 14:43:48 ####

speechd-0.6.7-10.3.i586 requires espeak, but this requirement cannot be provided

     ] Following actions will be done:
deinstallation of speechd-0.6.7-10.3.i586
downgrade of kttsd-4.4.0-2.3.i586 to kttsd-4.3.5-0.1.1.i586
install kttsd-4.3.5-0.1.1.i586 (with vendor change)
  obs://  -->  openSUSE
     ] keep espeak-1.39-3.3.i586

     ] break speechd by ignoring some of its dependencies

Do i need to log out and back into kde or something? I can’t do that at the moment as I have a backup running and few other processes I cannot interrupt.

EDIT: Forgot to say that that piping espeak to aplay works too.

downgrade of kttsd-4.4.0-2.3.i586 to kttsd-4.3.5-0.1.1.i586

Well if you’re going to use unsupported repos I guess you can troubleshoot it your self. This is an alpha/beta/milestone problem as far as I’m concerned.

But using the supported repos and festival it works fine for me, but I have to admit I just zyppered it into a clean room 11.2 vm install. But yes I needed to restart but that was because I got a load of kde bits updated.

So what are you saying? Because I’m using kde4.4.0, from the factory repo that I’m using an unsupported repo.

If that’s the problem then fair enough.

I’m saying if you’re using factory then you should be able to narrow the problem down.

I confirmed that with the released version you’re correct that espeak does no longer show as the manual for kttsmgr. Then also confirmed I still had additional problems with espeak needing to resort to piping into aplay which still seems a little wrong to me, but even if I wanted to I can’t bug report from a vm.

As for whether factory is actually the cause of your problem, I wouldn’t like to guess. What I do know is with OSS, Non-OSS, update and a couple of chromium repos it works fine.

OK, some progress. It seems that restarting the laptop has helped. This morning, the espeak synthesizer is showing up in the ktts manager so I’ve been able to set up a talker.

Now when I try some text to speech I get a message saying the speech-dispatcher is not configured or working. Hard to make out as the message is spoken and is not very clear. Sounds more like a 1980’s TTS.

Is there any reason why you don’t want to use festival?

OK mainly out of intrigue, how do you get espeak to show no matter what I do I can’t.

festival-1.96-122.2                           Thu 25 Feb 2010 12:12:40 GMT
portaudio-19-262.4                            Thu 25 Feb 2010 12:01:12 GMT
libjack0-0.116.2-5.1                          Thu 25 Feb 2010 12:01:10 GMT
espeak-1.39-3.3                               Thu 25 Feb 2010 11:56:04 GMT
kttsd-4.3.5-0.1.1                             Thu 25 Feb 2010 11:44:42 GMT
kdebase4-runtime-4.3.5-0.1.2                  Thu 25 Feb 2010 11:44:38 GMT
python-kdebase4-4.3.5-0.2.1                   Thu 25 Feb 2010 11:44:11 GMT
kdelibs4-4.3.5-0.1.1                          Thu 25 Feb 2010 11:44:07 GMT
python-kde4-4.3.5-0.1.4                       Thu 25 Feb 2010 11:43:57 GMT
kdelibs4-core-4.3.5-0.1.1                     Thu 25 Feb 2010 11:43:47 GMT
libkde4-4.3.5-0.1.1                           Thu 25 Feb 2010 11:43:43 GMT
oxygen-icon-theme-scalable-4.3.5-0.1.1        Thu 25 Feb 2010 11:43:27 GMT
kde4-filesystem-4.3.5-0.1.1                   Thu 25 Feb 2010 11:43:25 GMT
oxygen-icon-theme-4.3.5-0.1.1                 Thu 25 Feb 2010 11:43:09 GMT
libkdecore4-4.3.5-0.1.1                       Thu 25 Feb 2010 11:42:29 GMT
patterns-openSUSE-kde4-11.3-1.1.1             Wed 24 Feb 2010 12:13:10 GM

This is everything I have installed yesterday most are updates, portaudio I did testing since removing it still works. I don’t have speech-dispatcher but suspect it maybe needed for espeak.

Checking deps of kttsd and espeak I see no mention speech-dispatcher.

boo@linux-nuoc:~> rpm -q --requires kttsd
rpmlib(VersionedDependencies) <= 3.0.3-1                                                                                                       
kdebase4-runtime >= 4.3.5                                                                                                                      
kdebase4-runtime < 4.60.0                                                                                                                      
libqt4 >= 4.5.2                                                                                                                                
rpmlib(PayloadFilesHavePrefix) <= 4.0-1                                                                                                        
rpmlib(CompressedFileNames) <= 3.0.4-1                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 
rpmlib(PayloadIsLzma) <= 4.4.6-1
boo@linux-nuoc:~> rpm -q --requires espeak
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadIsLzma) <= 4.4.6-1

Checking with zypper verify my system verifies fine and also sounds like Stephen Hawkins

Is there any reason why you don’t want to use festival?
No reason other than festival did not show up in the ktts manager. Both festival and espeak were installed but only espeak showed up so I deleted festival.

Thinking about it now, I don’t remember exactly if espeak was pulled in when I installed ktts or not. I was messing around with various packages at the time.

Speech dispatcher is listed as the libspeechd module I believe

rpm -q --requires kttsd
rpmlib(VersionedDependencies) <= 3.0.3-1 
kdebase4-runtime >= 4.4.0                
libqt4-x11 >= 4.6.1                      
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadIsLzma) <= 4.4.6-1

rpm -q --requires espeak
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(PayloadIsLzma) <= 4.4.6-1

I didn’t really do anything special to get the espeak synthesizer showed up in the ktts manager. I installed ktts and espeak. then I assume you have to logout and back in (I restarted my PC only because I wasn’t able to logout and in at the time due to other processes running which I needed to complete).

The espeak just showed up in the ktts manager. The voice is terrible and for some reason it knocks my sound out after I try any TTS. It was mostly curiosity that got me to try TTS but I’ll probably leave it now until 11.3 comes out with KDE4.4.0 officially supported.

Out of interest can we use other voices like the AT&T voices with kttsd?

Interesting the opposite to what I have, though tbh I found espeak slightly better sounding. If your lucky I may upgrade to 4.4 seem strange to have an added dependency. Though it was only this thread that got me interested so don’t hold your breath.

In regards to voices not knowing anything about AT&T voices and little about TTS… Festival and espeak are the voice synthesizers, as far as I know you’re tweaking variables like phonemes to create the voice.

So unless AT&T voices are in a format that is like or similar to festival or espeak or any of the other voice synthesisers that kttsd can use I doubt it.

Festival Speech Synthesis System - 24 Voices

eSpeak: Voice Files

Edit looking at the kde page for kttsd it looks like speechd was added for 4.4

OK, if you decide to upgrade to KDE4.4 then report back here with your findings.

AT&T natural voices They sound very good in general. And looking at that site I see they now support Linux.

OK then, the AT&T voices will work but you’ll need to set up kttsd differently…

KDE Accessibility - KTTS - KDE Text-to-Speech System

Notice the IBM section this is what the AT&T voices are I think(Though if those are the prices that be a no go look at cepstral looks more affordable Cepstral Text-to-Speech ). Either way I suspect if not exactly the same a variant on using command would be the way to achieve it.

All a bit confusing but unless I’m mistaken using AT&T would need the command but I’m a little confused the KDE page talks about the IBM stuff and tbh those fonts(If they are sound fonts sound no better than festival) All in all cepstral looks the easier and cheaper option IMO.

TBH, I didn’t really look into the AT&T stuff too much. I know of it because my satnav used their voices. New or different voices could be purchased for ~$10 the last time I looked (a couple of years ago). I think those high fees mentioned on that site are the licensing fees to use the AT&T synthesizers in your product.

If the ktts is compatible with the AT&T voices then that would be great but like you say it’s very confusing. I’ll look at that Cepstral TTS.

My interest in this was purely curiosity so far.

Check this recent thread:

kttsmgr dosen’t work