This study examines the potential role of the auditory spectral integration in phonetic vowel quality decisions. Synthetic stimuli included a “virtual formant (F2)” which was produced by inserting two pairs of sine waves below and above a “perceptual target frequency” (the spectral center-of-gravity, COG). Intensity weighting across the pairs of sine waves created a virtual F2, i.e., an F2 percept which listeners formed although the formant was not physically present in the signal. Two different vowel series containing a virtual F2 were created by varying the intensity weightings of the sine wave pairs. The patterns of vowel identification decisions were similar with either the actual or virtual F2. The results are interpreted as evidence that the auditory system performs spectral integration across spectral components and can extract formant frequency information which in fact is not present in vowel spectrum