Lab: Measuring Stop-Consonant Voice Times
Download 15.35 Kb.
|
- Navigate this page:
- [lead + short lag]
- The lab exercise votsignals.wav Start ztool/SpeechTool and open ‘c:\ztool\votsignals.wav’. In some order
- Last question
| Lab: Measuring Stop-Consonant Voice Times
The purpose of this lab is to measure voice-onset times for stop consonants, but also to understand a little bit about how different languages divvy up the phonetic world in different ways to create their own, language-specific set of phonemic distinctions. The test signals for the lab are in three .wav files: ‘votsignals.wav’, ‘ba-pha.wav’, and ‘pit-spit.wav’, plus a spectrogram file: ‘ba-pha.fc’). Background VOT (voice-onset time) is the “burst-to-buzz” interval: the interval of time that elapses between the release of the occlusion (at the lips, alveolar ridge, or the velar/palatal area) and the onset of periodicity. The onset of periodicity can be seen on a BB gram by the 1st occurrence of approximately regularly spaced vertical lines corresponding to individual glottal pulses. Stops for which release and voicing onset are nearly simultaneous, or show a brief lag between release and voicing onset (or even simultaneous onset of release and voicing), are called short lag. Stops that show a significant delay (more than ~20-25 ms or so, but usually more than ~40 ms) between release and voicing onset are called long lag or voiceless aspirated. Positive numbers are used for these VOTs; e.g., +80 ms. Stops for which voicing onset precedes release are called lead or prevoiced, and negative numbers are used for these VOTs; e.g., -100 ms. Shown below in some order are signals and BB grams for three words: [pheso] (the Spanish word peso as it is commonly mispronounced by gringos), [beso] (the Spanish word for kiss), and [peso] (as close as I can get to the way it might be pronounced by a Spanish speaker). 
English has two phonological or phonemic voicing categories: voiced (V+) and unvoiced/voiceless (V-); i.e., /b d g/ contrast phonemically with. /p t k/. However, three phonetic (not phonemic/phonological) types occur in English. The three phonetic types are the ones mentioned above: (1) prevoiced/lead, (2) short lag, and (2) long lag/aspirated. How do you get two phonemic voicing categories out of three phonetic types? Easy. In English, lead and short-lag stops are allophones or allophonic variants (not variates) of /b d g/ (phoneme categories). They are in free variation, which simply means that speakers are free to use either lead or short lag for /b d g/ at their whim. English is a very permissive language in this way, and listeners are equally permissive – they treat lead and short-lag stops as equivalent and equally good instances of /b/ (or /d/ or /g/). In fact, fairly early in native-English speech development, listeners begin to have a difficult time hearing the difference between lead and short-lag allophones of /b d g/. (Language learning will do that sort of thing to a person, and this is one of the most important principles in understanding why 2nd language learning can be such a challenge, especially as a person starts getting up in years.) So, at the more abstract linguistic/phonological/phonemic level, English distinguishes (or contrasts) [lead + short lag] as a group with the long-lag category. Very important note: If you have forgotten or were never quite clear about the all-important distinction between the terms phonemic and phonetic – i.e., the difference between phonemic/phonological/linguistic categories on the one hand, and allophonic variants of those categories on the other hand – it would be a very good idea to dig out a phonetics text and review these concepts. There’s no way in the world to understand this material on VOT, and especially its role in 2nd language learning, without understanding those concepts. If you do not have a phonetics text, see if you can borrow one from a pal or see me. You can also find some discussion of these ideas here: http://homepages.wmich.edu/~hillenbr/204.html Look for the ‘language’ link (about the 3rd one down: Lecture notes: Language), especially the material starting around slide 66. (The slide numbers may change – search for the phrase Phonology and phonetics are not the same thing.) If you review this material and still don’t get it, see me. Back to the story: Spanish is different. Spanish also has two phonological or phonemic voicing categories: voiced and unvoiced, but these two voicing categories map onto phonetic categories in a way that is very different from English. Spanish has only two of the phonetic types that have been described: lead/prevoiced and short lag. The long-lag/aspirated phonetic type does not occur at all in Spanish. So how do you get two phonetic categories? Also easy: Spanish contrasts lead (phonologically voiced) with short lag (phonologically unvoiced). That’s pretty much it: the lead and short-lag phonetic types that are allophones of +voiced in English are distinct phonemic categories in Spanish. The lab exercise
Start ztool/SpeechTool and open ‘c:\ztool\votsignals.wav’. In some order you will see three words displayed: [pheso] (the Spanish word peso as it is commonly mispronounced by gringos), [beso] (the Spanish word for kiss), and [peso] (as it might be pronounced by a Spanish speaker). Bug the ‘bbgram’ button. Measure VOTs for each of the three words. You already know how to make time measurements from one of the earlier labs – the one in which you measured the fundamental period. Line the cursor up with the release burst, not the time (the ‘x’ value in the lower right), then line the cursor up with voicing onset and note the time, then subtract the two time values to get VOT. If you can see by eye that the two events are nearly simultaneous, it’s ok to just record a value of zero. That’s it. (If you don’t remember how to measure time intervals using SpeechTool, take a look at the earlier lab write-up.) Signal 1
Signal 2
Signal 3
If your previous displays are still on the screen, go to ‘Window’ and choose ‘Close all’ (or just close them one at a time). Open ‘c:\ztool\ba-pha.wav’, then click the BBgram button. (You’ll probably see a low-frequency buzz running through the entire spectrogram. Ignore it. You shouldn’t have any trouble differentiating the glottal buzz from the faint buzz that the laptop manufacturer put there.) Signal 1
Signal 2
If your previous displays are still on the screen, close them. Now open ‘c:\ztool\pit-spit.wav’. Bug the BBGram button. Signal 1
Signal 2
Last question: Anything odd going on here? I told you earlier that the English voicing system contrasts lead and short lag as a group with long lag/aspirated. Here we have two instances of /p/, so you’d think that they would both be long lag. Are they? If you made the measurements correctly, and I assume you did, they should both be aspirated. Are they? You should have found that one of them is long lag/aspirated ([phɪt]), but the other is short lag/unaspirated ([pɪt]). Now try this: Use the cursor to highlight just the /pɪt/ part of the sound wave (not the gram) for the signal on the left, then hit play. It will sound entirely ordinary; i.e., you’ll hear a normal aspirated /p/. Now try the same thing with the signal on the right – but once again, highlight only the /pɪt/ part of the sound wave, leaving off the [s] and the closure interval. Now hit the play button. Does it sound more like /b/ (or maybe the unaspirated [p] in [peso])? Now use the cursor to select spit, the whole word, including the /s/, then hit play. It should sound like a /p/ again. So finally, here’s the last question: What’s going on here? If you already know what’s going on from your previous journeys in the field, just tell me. If not, the answer is not too hard. The easiest way to figure it out is to take a look at ‘EnglishPhonologicalRulesControllingAspiration.pptx’, which is a small part of a phonetics lecture on consonant production. Note: There’s a long set up to this question, but only a sentence or two is needed to answer it. Download 15.35 Kb. Share with your friends:
|