To generate audio challange as an aid to the Text Captcha. We wanted to use one of the our voices, but then we thought it would be more attractive if you hear a woman speaking to you.
1) Cepstral Text to Speech Software is used to generate the audio clips. Website here.
A-Z, 0-9 and majority of the symbols can be read out. However, some like ` < > , . " : ( ) - cannot be interpreted by Cepstral.
So we gotta manually type in the english word, i.e for symbol &, we type ampersand. Guide follow from http://a4esl.org/q/h/9901/nr-keyboard.html.
The following audio are generated.
a) a - z (small caps)
b) A - Z (big caps)
c) 0 - 9 (numbers)
d) other symbols of the keyboard.
2) The audio clips are then edited and compressed to 32bits MONO using Blaze Media Pro, thereby reducing the size of each audio to abt 7-19kb. Website here.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment