[ruuk]

yes, you just can choose a language, no voice. I think google just has one voice per language available.

Thanks for confirming that.

And I also apreciate to have the google engine included in the tts addon, installing another addon to get it running would not be userfriendly.
I would go one step further and e.g. ask the rasbmc developer to pre install mpg123 already to overcome any dependencies outside xbmc (at least on the pi)

Perhaps it would be better to have the addon use apt to install it on user request. I assume they don't necessarily want to add to the size of the raspbmc installer.

By the way, could you add a volume setting to the google engine? or a global one to the tts addon? or would this also control the volume of xbmc? The google voice is a little bit loud here ;-) but I prefer a voice that is to loud instead of the ones you hardly can hear.

I've been thinking that a volume option should be added. It's now on my todo list. I'll add it for all the backends where the volume can be set.

Regarding alsa: I could not see any problems until now, it works really great on rasbmc.

Cool. Thanks for the feedback on that.

---

...aah forgot about the chache:

I think this is a great idea if it speeds up google tts. But I would not preinstall those soundfiles I would rather store them on the first request and just give the user two settings:
enable chache
and amount / files / size of the chache.

this way you would not have to set anything up and the system would get faster and faster if people activate it. And the traffic for downloading each sound only once cannot be that much :-)
Just my thoughts on that.
best regards,
Martin

This might be best. It would certainly be simpler.