Computer-aided conversion test system and method for generating intelligible speech

A computer-aided conversion test method for generating intelligible speech, comprising following steps of:

• step S1:
• providing an acoustic test module with a nasal-genio-oropharyngeal tract, the nasal-genio-oropharyngeal tract including a nasal cavity unit with two first openings communicating with external environment, an oral cavity unit with a second opening communicating with external environment, a pharyngeal unit connecting the nasal cavity unit and the oral cavity unit, and a laryngeal unit abutting another end of the pharyngeal unit;
• step S2:
• providing a central processing module, the central processing module including a transmitting-computing unit, a receiving-computing unit, a database storing a plurality of first phonetically oral cavity shape spectra, and a central computing unit configured to perform machine learning;
• step S3:
• disposing a transmitting module at one of the first openings of the nasal cavity unit and a first receiving module at the second opening of the oral cavity unit respectively;
• step S4:
• generating a detecting signal by the transmitting module which is controlled by the transmitting-computing unit, and receiving the detecting signal reflected by the nasal-genio-oropharyngeal tract by the first receiving module;
• step S5:
• converting the detecting signal received by the first receiving module into a second phonetically oral cavity shape spectrum corresponding by the receiving-computing unit;
• step S6:
• analyzing the plurality of first phonetically oral cavity shape spectra and the second phonetically oral cavity shape spectrum by the central computing unit;
• and
• step S7:
• determining whether the transmitting module or the first receiving module require being adjusted, if the central computing unit correctly comparing and identifying one of the plurality of first phonetically oral cavity shape spectra stored in the database which corresponds to the second phonetically oral cavity shape spectrum, no adjustment being required, if the central computing unit being unable to correctly compare and identify one of the plurality of first phonetically oral cavity shape spectra stored in the database which corresponds to the second phonetically oral cavity shape spectrum, an adjustment being required, and step 4 is performed again.