Figure 1. Speech wave, spectrogram, f0 (continuous line), and energy (dotted line) of original utterance er war mal mager "he used to be slim" (speaker KK, the author). Perceptually the uterance carries a medial peak, preceded by a high prehead; it is triggered by the slow f0 descent in the accented vowel of mager , strengthened by a concomitant rise in energy . Click on the figure for the audio.

Figure 2. Speech wave, spectrogram, f0 (continuous line), and energy (dotted line) of modified original medial-peak utterance er war mal mager "he used to be slim" (speaker KK, the author): the energy in the accented vowel of mager was reduced, which weakens medial-peak perception. This modified stimulus is the basis for the generation of the peak series. Click on the figure for the audio.

Figure 3. Speech wave, spectrogram, f0 (continuous line), and energy (dotted line) of original utterance er war mal mager "he used to be slim" with rising pitch (speaker KK, the author). This stimulus is the basis for the generation of the both valley series, low and high rising. Click on the figure for the audio.

Figure 4. Peak shift series in er war mal mager "he used to be slim". Spectrogram of modified original utterance and 10 peak contour shifts: peak maximum 60 ms before [m] - [a:] boundary, at boundary, and in 8 steps of 30 ms after the boundary. Click on figure for the audio presentation, which follows the sequence from left to right.

Figure 5. Valley shift series in er war mal mager "he used to be slim". Spectrogram of original utterance and 5 f0 valley contours: linear rise from [m] - [a:] boundary point, 3 steps of successive extension of low pitch scoop before the rise, and a linear rise starting from the boundary point at a higher f0 level. The series was generated with two f0 end points: 180 Hz and 300 Hz. The audio presentations follow the sequence as described. Click in the left section of the figure for the low-valley series, on the right for the high-valley series.

© Klaus J. Kohler, ipds Kiel, 2009