The main goal of the present experiment was to verify whether sensitivity to time interval changes in speech is greater than sensitivity to time interval changes between tones. The 12-syllable sentences were delivered in French (the main language of the participants) or in a foreign language (Slovenian) that was totally unfamiliar to the participants. Two types of sub-sequences were induced within the series of stimuli (sub-sequences in sentences or in tone sequences). Discrimination, as revealed by the Weber fraction in each condition, was much better in the tone conditions than in the speech conditions. Nevertheless, discrimination was excellent in all conditions: Weber fractions below 2% with tones and around 4.5% in the speech conditions. Moreover, the study revealed that familiarity with a particular language does not affect performance and that inducing internal sub-sequences within tone conditions, as opposed to using a series of equal intervals, does not lead to better discrimination. The fact that discrimination is better with tones than with sentences indicates that the extensive training provided by speech does not lead to higher sensitivity to time interval variations than simple tones do. Instead, it seems to indicate that the different acoustical variations in speech reduce the capability to discriminate time interval variations.