The present invention is a preprocessing apparatus including a voice input apparatus for acquiring an uttered voice, an analog-digital conversion apparatus for converting the acquired uttered voice to digital voice data, and a comparator for selecting voice data, having a level which is equal to or higher than a certain level, from the digital voice data and for outputting the selected voice data. The preprocessing apparatus also includes a voice data cutout apparatus capable of cutting out voice data having a level which is equal to or higher than a certain level output from the comparator, while taking a phoneme as a unit, and a voice data output apparatus for outputting voice data of the phoneme unit output from the voice data cutout apparatus.