As I have mentioned before, I run misterhouse on a very slow server. I also
have an Audrey and use it as a remote text-to-speech device. Now, when
misterhouse wants to say something, it calls my TTS engine to generate a
.wav file and then tells the Audrey to play it. The problem is that the TTS
engine takes around 10 seconds to generate the file, creating a pause in
misterhouse as it waits for the file.
I've enclosed patched versions of mh/code/common/audreyspeak.pl and
mh/lib/Voice_Text.pm that fix this problem by forking the TTS engine request
into a separate process, allowing misterhouse to continue on its way. I
have tested it and Jim Duda has also tested it (thanks Jim!), but we both
Are there any Windows users out there who have Audreys who could test the
enclosed to make sure that it works well? I would hate to commit this code
only to have it kill speech as I consider it a central misterhouse function.
Let me know if you can test and how it works for you. The enclosed file is
actually a zip, but has been renamed to .piz as sourceforge is currently