This study had the goal of determining if the amplitude of the surface electromyograph signals changes in terms of time of analysis and subjects, deaf or normal listeners, when estimated in a 250 ms of length window, visually determined, considering the most stable signal period from the center of the chewing cycle. In order to do this, groups with control subjects, listeners and deaf individuals, who made use of the Brazilian sign language (LIBRAS), were studied. All participants performed continuous 5 s of chewing for the electromyographic recording of the temporalis and masseter muscles. The normalized RMS values of three chewing cycles were compared between and among groups. The results from the Kruskall-Wallis test did not show any statistically significant differences (p > 0.05) between the normalized RMS values obtained in the three individual chewing cycles, for each of the two completed and evaluated cycles, in both groups studied. The Mann-Whitney test showed that the mean normalized RMS values obtained in the first chewing cycle were higher for the control group when compared to the mean amplitude values of the first chewing cycle of the group of deaf volunteers. It can be concluded that, in these experimental conditions, the RMS values obtained from the select windows of 250 ms length duration, in relatively stable periods of the electromyographic signal of chewing cycles did not suffer any changes in terms of EMG register duration, in both studied groups, but does give evidence of the differences among the groups.