In the present study, we tried to improve the discrimination of short temporal intervals marked by two brief visual signals with an extensive and massive training involving the discrimination of intervals marked by brief auditory signals. Two groups completed two sessions of visual interval discrimination (pre- and post-test). Between the two sessions, participants were either discriminating intervals marked by auditory signals (experimental group) or waiting for a period equivalent to the auditory training (control group). Once a method (called jackknife) is applied to reduce the statistical noise inherent in individual psychometric functions, the results show that visual duration discrimination is improved in the post-test portion of the experiment, but this effect applies to both groups. Therefore, it is not possible to argue that the gain is due to the auditory training. The discrimination threshold in the visual condition remained much higher than the threshold observed in the auditory mode.