by AdamW » Thu Sep 13, 2012 5:17 pm
hutmark we would love to develop that feature too!
At the moment we can take somebody reading a notecard worth of text and certainly turn it into other words and phrases, but we can't achieve the "say anything" quality of MI3. We usually record 10-20 hours worth of audio in a good TTS voice.
The question for us is Would people be interested in making their own text to speech voices if the recording times were shorter but the voices were more limited?