<html> <head> <meta http-equiv="content-type" content="text/html;charset=utf-16"> <title>Solution to Last Month's Mystery Spectrogram - Rob Hagiwara</title> <meta name="keywords" content="Hagiwara, phonetics, linguistics, spectrogram, spectrogram reading, mystery spectrogram, solution"> <meta name="description" content="Solution to last months mystery spectrogram"> <meta name="author" content="Rob Hagiwara"> <link href="../rzstyle.css" type="text/css" rel="stylesheet"> </head> <body bgcolor="#CCCCCC"> <h2>Solution for March 2008 (from October 2001)</h2> <div class="spg"> <p align="center"><img src="../spgs/sol0110.jpg"><br> <embed align="center" height="25" src="..wav/wav0110.wav" controls="playbutton" width="25" autostart="false"> <font size="+2">&quot;My phone rings too loudly.&quot;</font></p></div> <hr align="center" size="3" width="100%"> <p>The discussion below is minimally revised from the original (October 2001) solution. I've made some changes to the scale of spectrograms, as well having revised some of my transcription conventions since then. I've adjusted the paragraph headers to better match the transcription in the image, but I haven't changed the transcription otherwise--in particular, you'll notice that I used to segment aspiration away from stop closure. Idon't do that anymore.</p> <p>[m], IPA 114<br> Lower-case M <!-- bilabial nasal --><br> Nasals are always hard, and initial they're harder. You should notice that there's a 50 msec lowish-amplitude (compared to the following vowel) sonorant starting about 50 msec in. It has a nice low first formant, and decent formant at about 1100 Hz, and apparently a zero at about 1900 Hz. My nasal formant (F2) tends to be higher for [n] than [m], and even higher for [K]. So this being a relatively low nasal formant, the best guess is [m]. While not all the transitions in the following vowel are helpful, it's worth noticing that F4 and F5 both rise out of this consonant, a decent indicator of bilabiality. (extra points: why?)</p> <p>[a&#x026A;], IPA 304 + 319<br> Lower-case A + Small Capital I <!--low fronting diphthong--><br> Once you get past the transitions, F1 is about 800-900 Hz, F2 seems to have a very very short steady state at about 1200. About 50 msec in, the F2 is clearly rising, indicating fronting. The F1 remains moderately steady until the last 75 msec or so, when it starts to dive. The fact that it gets low before it starts to transition into the following fricative (trust me), suggests a separate F1 target that's lower than the first one. So we've got a low, backish vowel which is moving frontish and highish. There's only a couple of these in North American English.</p> <p>[f], IPA 128<br> Lower-case F <!--labiodental fricative --><br> This is fairly high amplitude for a fricative, but notice that it'snot appreciably darker (higher amplitude) in any particular band. It's not darker in the high frequencies (as [s] usually is) or centered near the F2/F3 range (as [&#x0283;] usually is). This is 'unfiltered', or fricative noise that isn't shaped by vocal tract resonances. That is, it doesn't echo around inside any vocal tract cavity. So the best guess is that this is extremely forward in the mouth, either [f] or [], The transitions into and out of (look at that F4!) this fricative are all wonky, so your guess is as good as mine. Well, I know what this one is supposed to be, so I have an advantage.</p> <p>[o], IPA 307<br> Lower-case O <!-- cardinal 7 --><br> Is this a diphthong or not? What the heck is going on with the transitions? Well, the F1 is sort of mid (around 500 Hz) for a lot of this vowel, and the F2 is quite low (starting around 1200 and dropping over the course of this vowel. So we're starting with something mid and back and moving higher and backer. Sounds like a pretty good /o/ to me.</p> <p>[n], IPA 116<br> Lower-case N <!-- dent/alv nasal --><br> Okay, forget everything I said about nasal formants. Because this doesn't even really look nasal (no zero). The Pole seems to be in the same frequency as the first one, and the transitions are pulled out of wahck by tthe following sound. Okay, it's sonorant, but with less amplitude than either adjacent vowel, so you know it's a consonant, or at least an onset. And the amplitude change, not to mention the F3 dropping out all at once, are sudden, punctual events, rather than ramped, so this looks like a good nasal (although depending on what kind of approximant you're looking at, this might not be the best kind of cue to be using. Mark it as a Nasal and come back later and test the hypothesis.</p> <p>[&#x0279;], IPA 151<br> Turned R <!-- central coronal approx --><br> Okay, F3 below 2000 Hz (barely), F2 pushed down by F3, F1 really low. Note the (short) steady state.</p> <p>[&#x026A;], IPA 319<br> Small Capital I <!-- betw card 1 and 2, "kit" vowel --><br> This is hard to guess, because the extreme deformation of F3 and F2 required in the preceding /r/ means a whole lot of movement has to take place before the targets in this vowel can be 'expressed'. The fact that both rise suggest movement toward a front vowel. The fact that F2 rises as much as it does suggests a very front vowel. The F1 here is lowish, indicating overall a high vowel. So [i] and [I] are the best candidates here. The relatively shortness of the vowel suggests [I] but it could be reduced from something else.</p> <p>[K], IPA 119<br> Eng <!-- velar nasal --><br> So getting back to nasals, you should recognize this one immediately. The zero(es in this case), the abrupt edges, etc. Note the high pole (formant), and how F2 and F3 both seem to head toward it. Velar &quot;pinch&quot; (F2 and F3 targeting the same frequency), and the height of the pole suggests a velar nasal.</p> <p>[z&#x0325;], IPA 133 + 402<br>Lower-case Z + Under-ring <!-- vd alv fric --><br> I'm not sure how voiced this really is. I remember trying hard to maintain voicing, but I don't think I did. But this is quite short for a good /s/. Call it a lenis /s/ or a fortis /z/. Dunno. Fricative, clearly, with highest amplitudes in the highest (visible) frequencies, well above 4000 Hz. Must be an /s/ or a /z/.</p> <p>[t], IPA 103<br> Lower-case T <!-- dent/alv plosive --><br> Quite short for a stop, but not quite short enough to be a flap. You might be tempted to argue for something else. Not the really really long aspiration afterwards. Clearly voiceless during closure and aspirated. Aspiration is centered quite high, like /s/, suggesting the release of an alveolar closure.</p> <p>[&#x02B0;], IPA 404<br> Right Superscript H <!-- aspirated release --><br> I've been fiddling with my IPA gifs, and a realized that the aspiration sign is technically a diacritic. I've been treating it as a separate symbol, because it corresponds to a segmentable bit of the spectrogram. (I'd ask the IPA for a ruling on this question, but I'm sure I wouldn't like the answer.) Noise, shaped sort of like /s/, but notice that the noise near the bottom is shaped like the begginning of the F2 that comes out of it, suggesting, that the noise is being shaped by the (I guess it would be the) front cavity 'behind' the alveolar constriction. /h/ noise is always shaped by the vocal tract resonances.</p> <p>[u], IPA 308<br> Lower-case U <!-- cardinal 8 --><br> I had a fight with someone about the status of the [ju] diphthong a couple of years ago. But this is an /u/ (low F1, so high vowel, low F2, so back and round) after an alveolar. Wells (1982) explicitly lists 'too' as a 'goose' vowel, but doesn't say in the section on the 'goose' set that it is definitely a [ju] word as opposed to an [u] word. I expect it's not, since the reflexes of /ju/ are usually not spelled with 'oo'. But in any case, you can see the merger of [u] and [ju] after alveolars in this form.</p> <p>[&#x026B;], IPA 209<br> Tilde L (Dark L) <!-- velarized l --><br> Interestingly, we have a clear discontinuity in the formants, i.e. a brief blip where the formants seem to transition and then suddenly reappear where they were. That is, we've got something here that has the same F1 and F2, if a slightly raised F3, as the preceding vowel. But the blip tells us that something must have changed, and this probably suggests that we've moved into a new phone. Check out the incredible lift F4 gets. /l/s in English, aside from being dark (low F2), tend to have a high F4.</p> <p>[a&#x026A;], IPA 304 + 321<br> Lower-case A + Upsilon <!--low fronting diphthong--><br> This isn't much of a diphthong, since vowel seems to stay low (high F1) more or less throughout this vowel. But the lowish F2 of /a/ (or possibly /Q/ lowers similalry to the way it did in [o] above. So start low and back (as opposed to mid and back in [o]) and then add some rounding toward the end.</p> <p>[d], IPA 104<br> Lower-case D <!--coronal stop --><br> This is a cruddy stop. It looks like a fricative. It's I expected this to look, if anything, like a cruddy fla. But this looks pretty long, far too long to be a flap. But it has al lt his noisy crap running around in it. The slight lift in F3 (and I can convince myself in F2) in the preceding vowel suggest an alveolar transition, but what else you're supposed to be able to get out of this is beyond me.</p> <p>[l], IPA 155<br> Lower-case L <!-- alv lateral (plain) --><br> Another /l/. This one looks even more like a [w] than the last one. Luckily this one is in a reduced syllable and I can tell you to ignore it....</p> <p>[i], IPA 301<br> Lower-case I <!-- cardinal 1 --><br> Low F1, so high. High F2, so front. There is a school of thought that word final high front unstressed vowels must always be [I]. But I think i say [i] in 'city' ([s&#x026A;&#x027E;i]. Anyway, high and front ought to be enough to work out what this is, if you have the rest of the sentence in place. </p> <p> <script type="text/javascript" language="JavaScript"> <!-- document.write(" Last modified: " + document.lastModified) // --> </script> </p> </p></p> <div class="footer"> <hr> <table width="100%"> <tr> <td width="25%"><span class="smallprint"><strong>Robert Hagiwara, Ph.D.</strong><br> Dept. of Linguistics<br> University of Manitoba<br> Winnipeg, Manitoba<br>CANADAR3T 5V5</span></td> <td width="50%"><span class="smallprint"> <center> <a href="http://home.cc.umanitoba.ca/~robh/">Current Mystery</a> - <a href="http://home.cc.umanitoba.ca/~robh/cursol.html">Solution</a> - <a href="http://home.cc.umanitoba.ca/~robh/archive.html">Past Mysteries</a><br> <a href="http://home.cc.umanitoba.ca/~robh/howto.html">How To</a> - <a href="http://home.cc.umanitoba.ca/~robh/research.html">Research</a> - <a href="http://home.cc.umanitoba.ca/~robh/courses.html">Courses</a><br> <a href="http://umanitoba.ca/faculties/arts/departments/linguistics/lab/">To the Lab</a> - <a href="http://umanitoba.ca/faculties/arts/departments/linguistics/">To the Department</a> - <a href="http://umanitoba.ca/">To the University</a></center> </span></td> <td width="25%"><img src="../images/bribbon.gif" alt="Support Free Speech" align="right"></td> </tr> </table> </div> </body> </html>