摘要 |
PURPOSE:To perform a reproduction which matches with a utilization configuration of symbols by providing a means which classifies various symbols of an inputted text and a means which transforms the symbols, that are better not to transform into a reading, to characteristic acoustic signals. CONSTITUTION:A symbol classification means 3 accurately classifies the symbols that are required to be accurately read like academic symbols and unit symbols among the symbols which are extracted from the text inputted by a text input means 1 through a symbols extracting means 2. A symbol-reading conversion means 4 performs the conversion of the symbols, which are classified as the ones that are better to be accurately read by a symbol classification means 3, to their reading. A voice waveform generating means 6 generates voice waveforms in accordance with the readings converted by the means 4. Moreover, a voice signal output means 7 outputs the voice waveforms generated by the means 6 as sound waves. A symbol-acoustic signal conversion means 5 converts the symbols, which are classified as the ones that are better not to be accurately read by the means 3, to acoustic signals which are beforehand determined. |