A method is described for compressing analogue speech waveforms to reduce the peak/mean ratio without adversely affecting the speech intelligibility or acceptability. The software is relatively simple and is suitable for implementation in a low-cost RISC microcontroller. Methods of testing the resultant intelligibility using a diagnostic ...
A method is described for compressing analogue speech waveforms to reduce the peak/mean ratio without adversely affecting the speech intelligibility or acceptability. The software is relatively simple and is suitable for implementation in a low-cost RISC microcontroller. Methods of testing the resultant intelligibility using a diagnostic rhyme test or phonetically-balanced world list are described.