en.nishimotz.com

text2wav is an experimental tool for galateatalk.

since : 2010-11-11

based on galatea_dialog_studio and jagtalk

http://hg.sourceforge.jp/view/galatea/dialog-project/file/02cc68132a0d/script/text2wav

• text2wav must be in the (project)/script directory
• character encoding for input / output is UTF-8.

$ ./script/text2wav 
Usage 1: echo '12345' | script/text2wav outputname
Usage 2: script/text2wav outputname input.txt

$ ./script/text2wav out utf8.txt 
$ ls out
out.log  out.log.eucjp	out.pros  out.pros.eucjp  out.runner_log  out.wav  out.wav.info  out.wav.info.eucjp

$ head out/out.log
* text
私は音声合成器です。
* arranged_text
私は音声合成器です。
* chasen result
<S>
  <AP orth="私は" pron="ワタクシワ" aType="4" silence="NON">
    <W2 orth="私" pos="代名詞" pron="ワタクシ" aType="0" aConType="">
      <W1 orth="私" kana="ワタクシ" pron="ワタクシ" pos="代名詞" orthBase="私" kanaBase="ワタクシ" pronBase="ワタクシ" lForm="ワタクシ" lemma="私-代名詞" form="ワタクシ" goshu="和" aType="0"/>
    </W2>

$ head out/out.pros
input_text: 私は音声合成器です。
spoken_text: 私は音声合成器です
number_of_phonemes: 29
total_duration: 1965
-----
sil [10]
w [40]
a [75]
t [85]
a [55]

$ head out/out.runner_log 
stdout:
stderr:
rep Run = LIVE
rep Speaker = female01
rep Speak.stat = PROCESSING
rep Text.text = 私は音声合成器です。
rep Speak.text = 私は音声合成器です
rep Text.pho = sil[10] w[40] a[75] t[85] a[55] k[75] U[40] sh[120] i[40] w[55] a[75] o[110] N[45] s[95] e[65] e[90] g[60] o[90] o[70] s[105] e[65] e[70] k[105] i[40] d[50] e[75] s[105] U[45] sil[10]
rep Speak.pho = sil[10] w[40] a[75] t[85] a[55] k[75] U[40] sh[120] i[40] w[55] a[75] o[110] N[45] s[95] e[65] e[90] g[60] o[90] o[70] s[105] e[65] e[70] k[105] i[40] d[50] e[75] s[105] U[45] sil[10]
rep Text.dur = 1965

$ head out/out.wav.info
input_text: 私は音声合成器です。
spoken_text: 私は音声合成器です
-----
sil [10]
w [40]
a [75]
t [85]
a [55]
k [75]
U [40]