Model Talker converts English text to speech by using a computer and the natural voice of the person recording to imitate natural speech as closely as possible. Model Talker is one of a new generation of software applications that is designed to be especially helpful to those who are losing their voices yet want to retain the qualities of natural speech. The company was testing its beta version in 2008.
Once available on the open market, Model Talker promises to revolutionize the way those with speech-inhibiting diseases, such as ALS, communicate. It will offer another alternative as they try to communicate with the outside world and it will do so in their natural voice. The only limitation to this is that the Model Talker TTS system can only work if the person makes their initial recordings while their voice is still in good condition.
Those wishing to set up a synthesized voice with the Model Talker will be required to record a number of hours of audio. A number of common phrases will be recorded, but the system will also mix and match different vocal sounds to create words. These individual sounds are called phonemes. Even if an exact phrase is not recorded, the text to phoneme feature can help create many more words than the actual ones being recorded.
Though there are other voice synthesizers on the market, the big advantage with the Model Talker is that the user's voice serves as a model. For some, their voice is an important part of their identity and they would like to keep that part of them intact. Model Talker is the pioneer in allowing users to do that.
The Model Talker synthesized voice can be used on the computer or with a speech generating device supported by the SAPI 5.1 standard. At some point, the product could even be available for smartphones with the Windows Mobile system.
While the software has made huge advances in text to speech systems, the Model Talker does have some distinct limitations. The most noticeable one may be in the voice quality itself. While the synthesized voice may have the user's particular sound, it will still sound somewhat awkward and robotic. Model Talker acknowledges there are more natural sounding synthetic voices in use, but notes these often take many more hours of recording under optimum studio conditions. The company claims its software is just as sophisticated as the larger commercial products.