SAPI 4

[Guy Rabiller]

Hi,

I've recently bought NextUp Talker, and so far I'm very pleased using it.
I'm using AT&T Natural Voices ( TTS Engine 1.4 )

I've been surprised, in good, noticing that SAPI 4 tags are accepted and correctly transmitted to the engine.

This is not mentioned in the doc, it might be a good topic to add, because this feature opens a all new world of speech synthesis from your application.

I'm using it for 'artistical' content creation, so I understand this has little to do with physical limitations or conversation tools, but I just wanted to let you know there are some other practical applications, and in those contexts, the SAPI 4 support ( or any markup language support ) is unvaluable for 'artistical' voice and speech synthesizing control.

I've encountered an issue however, regarding the "Advanced Editing" Tool from the "Pronunciation Editor". Despite the fact it is mentioned to work with the TTS engine, it doesnt ( or I didnt find a way to make it work ), and no AT&T voices are available in the "Test Voice" drop down menu.

This is a minor issue, because the same functionality is available externaly through the WinDictEdit tool from AT&T, so we can directly modify the used dictionary(ies) ( you can use several dictionaries at the same time by editing the tts.cfg file), and as soon as we change the voice in NextUp Talker ( from Mike to Crystal for instance ), modifications are taken into account.

Another minor issue, pause between words inferior to 0.1 ( it seems ) are not taken into account. Again, this is a minor issue, thanks to the SAPI 4 support.

I do have a suggestion however: Adding the concept of "libraries" for the Predifined Phrases, so we can save/load "libraries" of predifined phrases.

That's it. Thanks for this great application!

Guy.

[Jim Bretti]

Guy,

Thanks for the all the feedback.

Out of curiosity, which sapi4 tags are you using?

Here is what's happening with the advanced pronunciation editor. The advanced editor only works when using the Sapi 5 interface to the voice engines, and by default, AT&T uses the Sapi4 interface. To change this, display the Voice Options panel (under the Options menu) and switch the Sapi Version to "Use Sapi5 Only". Using the WinDictEdit utility as an alternative works just as well though.

Are you saying that between word pauses less than 0.1 seconds aren't working?

Creating 'libraries' of predefined phrases sounds like a really good idea. We'll look into implementing that.

[dasadler]

Just curious... was there ever any follow up on the idea of creating 'libraries' of predefined phrases? If not by the company, any chance of implementing (if not already present) the ability to import/export libraries created by users?

I would imagine there are a lot of useful 'libraries' of predefined phrases already developed by the user community.