[Tiptoi] Text-to-speech

[Ulrich Sibiller]

Hello,

to make experimenting a bit easier and save some time I played around
with text-to-speech software. This is a little example how to use
text-to-speech software to generate your wavs using Ubuntu:

1. sudo apt-get install libttspico-utils
2. pico2wave -l en-GB -w cool.wav "tiptoi is cool"

-> you'll find cool.wav which you can listen to with play.

For conversion to .ogg I found dir2ogg:

1. dir2ogg *.wav
2. rm *.wav (tttool will complain if it finds both)


For the okto1 example yaml file (see below) you can this:

--------------------------------------------------------------
#!/bin/bash
for i in homer marge lisa bart grandpa; do
   pico2wave -l en-US -w $i.wav $i
done

pico2wave -l de-DE -w hallo.wav hallo
pico2wave -l de-DE -w Richtig.wav richtig
pico2wave -l de-DE -w falsch.wav falsch
pico2wave -l de-DE -w WoBarnius.wav "Wo ist Barnius"
--------------------------------------------------------------


$ cat okto1.yaml
--------------------------------------------------------------
product-id: 42
comment: test
init: $mode:=0

welcome: hallo

scripts:
  8066:
  - $mode==0? $mode:=1 P(WoBarnius)
  - $mode==1? $mode:=0 P(grandpa)
  8067:
  - $mode==0? P(homer)
  - $mode==1? P(Falsch)
  8068:
  - $mode==0? P(marge)
  - $mode==1? P(Richtig)
  8069:
  - $mode==0? P(bart)
  - $mode==1? P(Falsch)
  8070:
  - $mode==0? P(lisa)
  - $mode==1? P(Falsch)
--------------------------------------------------------------

@Joachim: maybe an additonal switch for the assemble command could
also generate this on the fly. This way you could hold all your text
in the yaml file without maintaining any wav/mp3/oggs. Perfect during
testing/development. After you are satisfied with your yaml you cold
still create better sound files.

Uli