The effect of low-pass filtering of the digital frequencies in quantized speech
Edward Michael O'Brien, David J. Pogue · Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society · 1988
Monosyllabic words were sampled (time-quantized) at a rate of 20 kHz. The signal was then converted into a trilevel waveform, that is, amplitude-quantized to either a 5 V, -5V, or a 0 V (squelch) level. Digital low-pass filtering was accomplished by reducing any positive or negative wave to 0 V if it has N or fewer quantizing intervals (50 mu s) in it. The effects of three low-pass filtering conditions (N equal to 1, 2, and 3) were compared to those of no filtering. The percentage of words correctly perceived was 29% for the no-filtering condition and 21%, 15%, and 9%, respectively for N=1, 2, and 3.>