Re: Is there a project that is jovie successor? (fwd)

Thanks, interesting.  I see there is a Debian package but I looked and 
can't find anything for opensuse or in rpm form.  I searched for larynx 
and linux and found a very informative thread on the state of tts and 
linux.  It looks like he is trying to get it to work with 
speech-dispatcher, so maybe it will be something useful before too long.

https://news.ycombinator.com/item?id=28651588



On 10/29/21 7:46 PM, Nickolay Shmyrev wrote:
> Hello guys.
>
> You need to checkout this modern technology one day:
>
> https://github.com/rhasspy/larynx <https://github.com/rhasspy/larynx>
>
>> 29 окт. 2021 г., в 20:12, Gustav Degreef <[email protected] 
>> <mailto:[email protected]>> написал(а):
>>
>> No problem about not replying sooner.    That West Indies voice is 
>> exactly what I hear also.  It is output at the same time the regular 
>> espeak English voice, so the output is incomprehensible .   I had 
>> said Scottish sounding, but you are more correct.
>>
>> No I was not able to get the "hello_speak example" to work. I can't 
>> get the program to build.   I have been trying to figure out why my 
>> test opensuse 15.3 system does not have the problem of the duplicate 
>> voices and the 15.3 system which I need for work does have the 
>> problem.  It seems that when I was able to install Jovie on the 15.3 
>> test system it was because it came from an unsupported repo which 
>> installed some Qt5 Frameworks packages, but I can't figure out which 
>> ones.  And now that I look again that unsupported repo no longer has 
>> Jovie available.  Not sure where to go from here.
>>
>> Jeremy, thank you for all your patient efforts and for undertaking to 
>> improve the situation.  I will keep an eye out for the improved 
>> Kmouth.  Gustav
>>
>> On 10/29/21 5:48 PM, Jeremy Whiting wrote:
>>> Gustav,
>>>
>>> Sorry, I didn't respond yesterday. The espeak.log you sent looks 
>>> like espeak is getting text fine from kmouth and okular. I'm betting 
>>> that since kmouth and okular do not tell QtSpeech which voice to use 
>>> it's using some default that isn't ideal. Were you ever able to get 
>>> hello_speak example to work? With it you can select each voice that 
>>> QtSpeech knows about and probably see what default it's using for a 
>>> given language. Here when I open it on my opensuse 15.2 vm it starts 
>>> with English, but has some west indies voice as the default which 
>>> pronounces everything strangely. If I change it to another english 
>>> voice it all sounds the same as spd-say.
>>>
>>> So it won't be in the next release since feature freeze will be soon 
>>> or has already happened, but I'll get voice selection in at least 
>>> okular's settings window where you can already select between flite 
>>> and speechd. And do likewise for KMouth with the other improvements 
>>> I'll do to KMouth. I think once you're able to select a voice (and 
>>> language) in those just like in hello_speak it should be pretty 
>>> straight forward to make it sound the way you want.
>>>
>>> BR,
>>> Jeremy
>