[PATCH 4/7] replace EVoiceType with SPDVoiceType

[Hynek Hanke]

On 5.11.2010 14:01, Tomas Cerha wrote:
> The predefined symolic voice names were meant as a sort of abstraction.  The 
> idea was
> that the user would be able to map them to available voices without requiring 
> the ATs to
> deal with that.

I think both concepts have proven useful for different purposes. The same
sort of abstraction is used in SSML as well, see:
     http://www.w3.org/TR/speech-synthesis/#S3.2.1

However, as other people proposed, I was also thinking the two
approaches can be joined into one:

1) TTS systems reporting available voices and their attributes
as described here:
     http://cvs.freebsoft.org/doc/tts-api/tts-api.html#Voice-Discovery
and here
     
http://cvs.freebsoft.org/doc/tts-api-provider/tts-api-provider.html#Voice-Discovery

2) setting just synthesis_voice
     http://cvs.freebsoft.org/doc/tts-api/tts-api.html#Voice-Selection
     
http://cvs.freebsoft.org/doc/tts-api-provider/tts-api-provider.html#Voice-Selection-_0028text-protocol_0029


I think we still need to keep the old SSIP SET VOICE for backwards 
compatibility
(and deprecate it), but this is not any problem, because its mapping to 
set_voice_by_properties()
is very straightforward.

Best regards,
Hynek Hanke