Solution for October 2004


"Things'll be fine by next week."

Theta
[θ], IPA 130
Well, that's just the loudest fricative I've seen in a long time. If it were more 'dense' it would look more [ʃ]-like. I'd also like to see more in the way of a clear 'zero' below the 1500 Hz mark or so. So [ʃ] is probably not an option. But it's definitely fricative. And being initial, I suppose I'd expect it to be particularly strong, so if this were weaker, it would look more like a non-sibilant. If that makes any sense, which it doesn't, since I might just as easily argue the opposite direction the next time. So anyway, Going for a fricative. It's not [h]-like, since that would definitely not have the zero in F1. It's obviously voiceless, so that leaves [f, θ], or [s]. Which is not saying a great deal. [s]s have been remarkably low-amplitude of late, so I'm not sure what to tell you here. The formant shaping, especially F3--I guess--suggests something that isn't [s], since it suggests narrower band(s) than the single broad band of [s]. Also, seems pretty flat, amplitude-wise, without really getting much stronger in the higher frequencies (by which I mean above 4000 Hz). So probably [f], or [θ]. But exactly which I'm not sure I could say.

Small Capital I
[ɪ], IPA 319
Well, this at least is easy. The F1 is low of the middle-frequency range (around 500 Hz), so this must be high. The F2 is way high, rising up from just below to just above 2000 Hz, so this must be way front. And it's really, really short. (Q for beginniers: What does the duration of this vowel have to do with anything?)

Eng
[ŋ], IPA 119
Well, this is fully voiced, but except for the F1-region, there's much less energy than in the vowel. There's also a funny discontinuity just before 400 msec that I'm justgoing to ignore, bbecause I'm not sure what it is. But if you think it's a separate sound, you also have to fit it into a word or phrase. So anyway, we've got something abruptly dampening the amplitude, but fully voiced and apparently sonorant. Which suggests a nasal. The F2 and F3 transitions in the preceding vowel look 'pinchy', suggesting a velar (front velar, given the F2 frequency). The concentration of energy in the F2-F3 range and the zero above give this a distinctly [ŋ]y feel.

Lower-Case Z
[z], IPA 133
Well, when this thing gets fully noisy, the noise is broad band and high frequency, i.e. [s]-like. That said, it's also very short in duration, and apparently fully voiced. Which is [z]-like. Now, about that funny F2 thing or whatever it is starting at about 400 msec and going on at elast until the middle of the following sound. That might be transition from the release of the velar closure to 'fully slitted' fricative. That is, I don't think I released the nasal into a fricative, but into a tongueshape moving *toward* the fricative. Or at least the center of the fricative. TMSAISTI.

Tilde L (Dark L) + Syllabicity Mark
[ɫ̩], IPA 209 + 431
Ah, finally something that looks like it's supposed to. Well, sort of. It looks like a vowel, but it's very weird for a vowel. Look at that F1. Mid to vaguely high. Look at that F2. Back and/or round. Very back and/or round Look at that F3! Up around 2800 Hz or something. That's about as high as F3 ever gets (unless it's coupled with the high F1, more on that later). High F3s are a correlate of [l]. I'm still waiting for Frank Guenther or Carol Espy-Wilson to explain to me exactly why, but they are. The extremely low F2 is a sure sign of velarization, or backing, or 'darkness' to use the traditional term. It's loud and vowel like because it is, at least functionally, a vowel.

Lower-Case B
[b], IPA 102
This is technically short enough to be a flap, I guess, and I've produced very flappy-looking [b]s, but since a dynamic contact-type approximant is not recognized with a seperate symbol by the IPA, and frankly this might just be long enough to count as a real stop, I'll just call it a [b]. Note the sharply down- pointing F3 and F4 in the preceding sound, and the rising F4 into the following vowel. Whichis really the only good evidence of anything in particular. There's no evidence of velarity or coronality really, which leaves labiality as the best choice. Fully voiced, potentially approximant or fricative, but just stoppy enough to probably count.

Lower-Case I
[i], IPA 301
Ignoring the F2 transition for a moment, safe because of both the extremely low F2 starting position contributed by the dark /l/ and not helped at all by the bilabial, the 'extremum', or turning point, or whatever you want to call it for the F2 is very, very high. And since this vowel is considerably longer, I'd say this is a good candidate for /i/.

Lower-Case F
[f], IPA 128
Ah, another of these fricatives. THis one looks even more like an [s] than the preceding fricatives, but it's a little weak to be an [s] that is that long. Even the noise in the [z] is louder. So this probably isn't a sibiliant. Certainly not voiced. The absence of a clear F1 and for that matter the absences of a *lot* of formant-like filtering rules out [h]. So again we're left with the labiodental and (inter)dental fricatives. And there's not a lot of useful information in the transitions. I suppose that F3 rising out of it is as good as the F3 falling out of the first one, but that's not saying a great deal. So how do you tell the difference? If you figure that you, you can write a dissertation on it that will be cited *forever*.

Script A + Small Capital
[ɑɪ], IPA 305 + 319
The reason that the F3 transition isn't really useful in terms of the transition from the preceding fricative is that it looks like the F3 is coupled with the F1. There's really no other reason for it to be this high, but I've been looking at a fair number of open vowels lately, and a lot of them have F1/F2s that are mirrored very closely by their F3/F4s. And I think it was the Maeda model that allowed F3 to just be next natural resonance of the F1 and the F4 to be the octave (I guess that's what I'm saying) of the F2. So I think that's why the F3 here is so *freaking* high. Meanwhile, the F1 is very high, so this vowel is about as low as it can get, and the F2 is about as low as it can get, so this must be back and/or round. Then at about 900 msec the F1 starts to fall and the F2 starts to rise. So this looks like an /aj/ diphthong. But since my transcriptions are supposed to be about the acoustics, and there's no way to read that nucleus as [a], I've tried to be literal.

Lower-Case N
[n], IPA 116
Ah, fully voiced, resonant, but definitely less open than a vowel. Nice and flat too. Classic nasal material. The poles are in the F1 range, just below 1500 Hz and up around 2600, which are pretty classic for my /n/. The transitions are *consistent* with coronal, but not amazingly helpful. The lowering F3 is worrying until it becomes clear that the F2 dives away from it. So that isn't pinch. That transition is really important.

Lower-Case B
[b], IPA 102
Ah, finally, a [b] that looks likea [b]. (I'm really tired, and I was really starting to despair these cues.) Nasal aside, we've got something quite voiced, but the nasal resonances seem to click off (more or less) at one time. The burst, if that's what that is, is very low frequency. All the transitions into the following vowel rise. Woo hoo! Clearest evidence for a nice, voiced [b] I have ever seen.

a

Lower-Case A + Small Capital I
[aɪ], IPA 304 + 319
Okay, now this looks like an [aɪ]. The F1 isn't quite as high as in the previous diphthong, and the F2 isn't nearly as low. So this nucleus might read better as an [a] than as an [ɑ]. Still the F2 rises, and though it's a little ambiguous I can convince myself the F1 is falling. So it's a diphthong.

Lower-Case N
[n], IPA 116
Another one of these.

Epsilon
[ɛ], IPA 303
The F2 here is ambiguous, but it's definitely higher than the high vowels we looked at early on in the utterance. So this ain't high. It's very front, so we're looking at something frontish and not high. Doesn't narrow it down much, but considering what's coming up, it'll probably have to do. It's not really long enough to be a reallo low vowel (even if it did have the F1, which it doesn't really). So think mid. So that leaves [e] and [ɛ]. Frankly, I'd have guessed [e] from the direction the F2 is taking, but it turns out that it's rising due to the upcoming velar (you can tell becasue the F3 is coming down the same way), so I'd have been wrong.

Lower-Case K
[k], IPA 109
Anice little stop, mostly voicels, with velar pinch into it and, well, possibly a little coming out (but that's a stretch, even fro me. Nice big low- frequency release transient or something, which makes it looka little bilabial, but then there's no explanation for the F2. So velar it is ...

Lower-Case S
[s], IPA 132
... released smack into the shortest [s] I've ever seen. But at least it looks like an [s], at least as far as its frequency/amplitude profile goes. But cut very short ...

Lower-Case T
[t], IPA 103
... by the shortest fully aspirated [t] I've ever seen. Okay, it has to be a [t] because nothing else releases like that, not even [d].

Lower Case W
[w], IPA 170
So then we have a very low F2. Note that it is not accompanied by a particularly low F3, nor a particularly high F3. So we've got a neutral F3 combined with a low F2, which is pretty classic for [w]. With a very fast transition into ...

Lower-Case I
[i], IPA 301
... a *very* front vowel. And probably quite high to book, givne that the F1 has none of the ambiguity of the previous mid vowel. So this is probably [i].

Lower-Case K + Superscript Lower-Case H
[kʰ], IPA 109 + 404
And there's that velar pinch again, followed by a nice long gap, and a fairly center-frequency heavy release transient. I have to remember to stop marking these things as aspirated when they're merely released. The noise is a little odd looking, being slightly [s]-shaped, in the sense of being heavy ikn the high frequencies, but noitice those bandwidths, rather than the one big band.