Figure 1.Frequency transitions (changes) on the front end of vowel-producing formants produce may consonant sounds.

These consonants must attached to formant because they are all voiced consonants: the vocal tract iss set in vibration when you say them (unlike /s/, for example, which is unvoiced -- the vocal cords don't vibrate). This is why they were written as /ba/, /pa/, etc., instead of /b/, /p/, etc. Each consonant has its own pattern of transitions(6). Some sets of consonants, like /b/, /d/, and /g/, differ in the slope for their transients. Other sets of consonants, like /ba/ and /pa/ differ in voice onset times(VOT), the time between the start of the consonant and the start of vibration in the vocal tract. For /ba/ the voicing starts when the speech sound starts; for /pa/ the voicing is delayed by about 65 milliseconds Figure 3 illustrates the difference in Voice Onset Time(VOT) for /ba/ and /pa/ in the 1st (lower) formant relative to the start of the transition of the 2nd format. Simultaneous onset produces /ba/. A delay of 65 milliseconds (65/1,000 second) produces /pa/.

Figure 2.Voice Onset Time (VOT) is the delay between the start of the speech sound and the beginning of vocal cord vibration. VOT distinguishes between /ba/ and /pa/, /ga/ and /ka/, etc. The first of each pair has delay in vocal cord vibration; the second normally has a delay of about 65 msec.

To return to the exercise, close this browser window.

To return to the exercise, click on the Back button in the upper left corner of the browser window.