[Kde-accessibility] dcop kttsd kspeech sayMessage mess talker

[Bill Haneman]

Mikolaj Machowski wrote:

>Dnia sobota, 1 stycznia 2005 20:16, Gary Cramblitt napisał:
>  
>
>>On Saturday 01 January 2005 11:07 am, Mikolaj Machowski wrote:
>>    
>>
>>>I have two talkers defined in kttsmgr:
>>>Id  Lang
>>>1   English
>>>2   Polish (default)
>>>How with dcop -
>>>dcop kttsd kspeech sayMessage 'message' talker
>>>use chosen one? Currently dcop uses always default talker.
>>>      
>>>
>>dcop kttsd kspeech sayMessage "Hello World" 'lang="en"'
>>
>>    
>>
>IMO this is not good solution. What if user has, for example, two voices
>for english, male and female and want to switch between them?
>  
>
It's fairly rare for a voice to support multiple langs, so perhaps tying 
the lang to the voice makes the most sense here.  The downside of this 
approach is that the client must decide to switch voices based on LANG.

Ideally, lang would be supported as part of a general markup approach, 
and the TTS service or wrapper would be responsible for voice 
substitution if the selected voice can't handle the specified LANG.  You 
might want to have a policy setting for whether such voice substitution 
was allowed.

- Bill

- Bill


>Wouldn't be better to use kttsmgr voice IDs?
>
>m.
>
>  
>