Figure 10.

Narrow-band spectrograms of versions of the VCV ∕ɑ k ɑ∕ as spoken by the female talker of experiment 3. The top panel shows the original speech token, followed by the conditions of particular interest (for details, see the text). Note the similarity between conditions 3s300 and Px. Condition Mx is similar to Px, but with a fixed F0, identical for all three channels, as can be seen from the fact that all the spectral components are fixed in frequency (unlike the varying ones in Px). Condition Dx also uses fixed F0s, but ones which vary from channel to channel (note that the spectral components in the highest frequency channel are more widely spaced than those in the lower two). Also of interest is the fact that the transient release burst of the ∕k∕, shown prominently by the dark “blobs” for both the speech and 3s300 sounds at consonantal onset, is more or less eliminated by the 30 Hz envelope filter used in the other conditions.