Solution to Last Month's Mystery Spectrogram

Solution for August 2008 (originally from December 2001)

"My feet get cold at night."

Minimally revised from the original. Wow, my style has changed. And not just in 'official' ways--like treating aspiration (properly) as a diacritic and not a separate thing.

[m], IPA 114
Lower-case M
It would have been easy to miss this consonant, but there are three or four good glottal pulses starting at about 75 msec telling you that there's *something* there. It's voiced, it's resonant, and it has a zero (actually several), suggesting a Nasal, and the pole near 1000 Hz suggests /m/ for my voice. /n/ has a higher pole, closer to 1500 Hz. But that may just be me.

[aɪ], IPA 304 + 319
Lower-case A + Small Capital I
More than anything else, I wanted my students (this spectrogram was offered as extra credit on their take-home final) to notice this pattern. F1 starts high and dropps (this is tough to see, but once you know it's there it's pretty obvious). F2 starts incredibly low, indicating backness, and goes very high. the combination is classic for the /ai/ diphthong, although the appearance of steady states is a matter of some dispute. There really aren't any here.

[f], IPA 128
Lower-case F
I've got to do something about my fricatives. This is centered too low to be a /s/, it doesn't have enough structure to be /h/. It sort of has the structure of a post-alveolar [ʃ], but it isn't really strong enough to be sibilant. That leaves the labidodental and the (inter)dental (for English voiceless fricatives in my normal speech), and your guess is as good as mine.

[i], IPA 301
Lower-case I
If I've said it once, I've said it a thousand times. Super-high F2, mostly low F1. So we're looking for the frontest vowel we can find. And it's relatively high.

[t], IPA 103
Lower-case T
I worked hard to make sure there was a release here to tell the relatively inexperienced reader there had to be two things in this gap. How exactly you tell that this is alveolar as oppose to anything else is beyond me. The burst looks double (although I think that's really the burst of the /t/ and the closure of the /k/, and the transitions are look labial (but who can tell when the F2 and F3 are that high in the vowel). All I can say is notice the energy in the (first) release. It has the 'slope' of an /s/. So there.

[k], IPA 109
Lower-case K
In fact, this is a front /k/, owing to the rontness of the following vowel. Note the F2 F3 pinch in the release, and the noise concentrated in the F2 F3 conjoined region. Fairly classic for velars. This is voiceless but unaspirated (in the chinchilla-VOT sense), so this is probably phonemically a /g/.

[ɛ], IPA 303
Epsilon
Vowel. I take it that isn't controversial. Check the F1. It's higherthan the /i/ precedign, and a lot lower than the /a/ at the begining of the /ai/ diphthong. So it's mid. Note teh F2. Though falling, it's still high, indicating front. Mid and front. If anything, it's centralizing rather than raising or fronting. So probably 'lax'. Review your vowel features and you're there.

[t], IPA 103
Lower-case T
Okay, this looks more like a flap, but this is so not a flapping environment. So either the flapping environment is wrong, or it's not a flap. I don't know the answer. The duration and amplitude of the 'release' noise suggests that a bit of pressure built up. So let's call it a stop. Note also the shape of the noise in the release, i.e. that it's stronger in the higher frequencies than otherwise. I.e. looks like /s/. So what stop-like thing is most likely to look like /s/?

[k], IPA 109
Lower-case K
Long gap, followed by a low-frequency double burst. Note also the transitions in the following aspiration. Not very pinchy, but in the pinchy directions.

[ʰ], IPA 404
Right Superscript H
Aspiration. Loong. (Review the aerodynamics of aspiration and place of articulation.) The double burst and the loooong aspiration suggest velar for the previous stop.

[o], IPA 307
Lower-case O
Note the F1, in the same range as the mid vowel previous. note the incredibly low F2. Not [u], due to the mid-ish F1. But very back and round. Must be [o].

[l], IPA 155
Lower-case L
It would have been easy to miss this , but post hoc I can convince myself that there's a change in bandwith of all the formants, but particulalry F3 and F4, about 2/3 through the 'vowel' here. The F3 and F4 take a teeny little jump here as well, but otherwise this looks quite back (and round). So we're looking for some kind of approximant which is back. This could be /w/ or /l/, if you catch it at all. Turns out to be /l/ (velarized), and the little jump in F3 and F4 (okay, mostly F4, but it's supposed to be in F3...) might help us decide it was /l/ and not /w/.

[d], IPA 104
Lower-case D
Another goofy stop. Short, and this one is arguably in a flapping environment. Okay, so maybe this one is really a flap. Sue me.

[æ], IPA 325
Ash
I swear I tried to make this a full vowel. But it's looks awfully schwa-like. It looks mid-ish, but it looks pretty non-descript F2 and F3 wise. Okay, call it a schwa and it won't matter in the long run. Turns out to be a function word anyway.

[t], IPA 103
Lower-case T
Stop, good release, not really useful in the transitions into it. Looks kind of labial. But it's a /t/. Sue me.

[n], IPA 116
Lower-case N
Fully voiced, resonant, nice little zero, F2 near the neutral range, which for me always means alveolar.

[aɪ], IPA 304 + 319
Lower-case A + Small Capital I
Another one. Same reasons, but this one has a better steady state in the 'a' portion.

[t], IPA 103
Lower-case T
Well, there's a buncha cues for this, but since it's a stop, they're mostly in the transitions into it and the aspiration out of it. Utterance final, so there's probably a low boundary tone floating around here, in addition to the general tendency for me to glottalize final /t/.

[ʰ], IPA 404
Right Superscript H
Okay aspiration. Note the shape of the release noise. Looks like an [s]. But the rest of the noise is organized into formants like a good /h/.