[EAS] Question re CAP and text-to-speech

[David Turnmire]

Its my understanding that the text-to-speech translation can happen at 
either end.  Doing so at the emergency management end would have some 
advantages, including: 1/ The speech-to-text system could be "trained" 
to know how to pronounce place names, etc and otherwise potentially do a 
better job than the typical decoder,  2/ save the EMs some time since we 
want the text for TV display as well as non-EAS messaging devices... as 
well as the need for voice for broadcast, all-hazards-radios, etc. and, 
3/ possibly improve sound quality over the sometimes poor microphones 
and mic techniques (and noisy rooms!) of the past.  The resulting voice 
could be sent as an attached file via CAP... and/or delivered via legacy 
analog means.

On the other hand, some decoders have the ability to generate voice 
automatically if there isn't a wave file included with the CAP message.  
That is better than nothing... but at least for the CAP box I have, the 
voice is distinctly "computer like".  I haven't explored it much, but 
I'm pretty confident it would have trouble accurately pronouncing 
certain words.  In any event, if you want it to provide a useful voice 
message, you need to have the text part... which basically means CAP.  
You could simulate it from just the SAME data, but then you have the 
same incomplete/misleading voice message as we have had for years as 
text on TV displays... no thanks!

Dave