Attached is a much better version of "ns04kismet" (source and binary). This has WAV processing if you use a two-space delimiter. For example, my /usr/local/etc/kismet_ui.conf has:
Code:
speech=true
festival=/home/raf/ns04kismet.exe
flite=true
speech_encrypted=%s is closed
speech_unencrypted=%s is open
(yes, the default strings are longer, but they took too much time for me with a dense area so I shortened them to match my NetStumbler strings).
You'll notice that it uses magic numbers of 3 off the end and 2 off the beginning. They are for the " )\n" and "( " in the passed input.
The only things that come to mind that could use improvement are
- checking for two speech strings within a single clock second; currently it just overwrites what was there with the new speech
- using registry calls natively rather than spawning regtool.exe frequently
If anyone is using festival/flite on a Linux/BSD/etc. box and wants WAV pronunciations, I could probably code up a script to do it. I'm thinking that, ideally, Kismet would just append to a specific file, and the script would read in the list of WAVs then "tail -f" the file, speaking/"play"ing as it went. Anyone interested?