Auditory analysis of synthetic vowels /a and æ/ was investigated by the pulsation threshold paradigm. Vowel maskers were alternated with sinusoids with 50% duty cycle and period of 250 ms. Listeners varied the level of the sinusoid to pulsation threshold or to detection threshold (simultaneous masking). The pulsation threshold data generally parallel the envelopes of the vowel spectra at moderate SPLs (50–60 dB). At high SPLs, the pulsation thresholds representing the formant regions shift upward for F2 and downward for F3. The troughs between formants are not as deep as those at lower SPLs. In some individuals they fill up so that formant regions are not evident. Increases in SPL then do not necessarily lead to increased suppression. Simultaneous masking (detection) fails to show anything but the first formant. It appears that “upward spread of masking” which has been used to explain simultaneous masking data of vowels is only a decrease in suppression effects.