Solution to Last Month's Mystery Spectrogram

Solution for November 2006

labeled spectrogram
"Most require careful management."

Woo, a rough one. Very. Things will get easier in January. Probably.

Lower-case M
[m], IPA 114
From about 50-150 msec there's strong voicing, a weak, flat formant at 1200 Hz, another around 2600 Hz and one higher than that too, but nothing in between. So something with that kind of weak resonance/zero structure, and flat, has to be some kind of a nasal. The pole at 1200 is lower than I'd usually get for an alveolar, though a big higher than I'd normally get for a bilabial. But the transition in the following vowel is in no way alveolar-looking, so there you go. Probably in initial position my tongue wasn't as low as it might have been, effectively shortening the side cavity. Think about it.

Lower-case O
[n], IPA 307
F1 at about 500 Hz, F2 just above 1000 Hz. Don't ask me why the F3 is high. But something that's got a mid-vowel F1, a back/round vowel F2. And basially flat, rather than obviously diphthongized.

Lower-case S
[s], IPA 132
Not the strongest I've seen, but whatever. Probably the effect of the syllabic psoition (coda, but not final). A single broad band of noise, centered fairly high, and without a sharp drop off in the low frequencies. Don't be distracted by the weak perseverative voicing.

Lower-case T
[t], IPA 103
Short gap, followed by [s]-shaped release noise (pulled down a bit in frequency by coarticulation with the following sound). So alveolar [t] has [s]-shaped release noise. Discuss.

Turned R
[ɹ], IPA 151
Hard to tell, but that's F3 just above teh F2. That thing up around 3000 Hz is just too high to be F3. So the F2 starts about 1100 Hz and rises, and F3 starts around 1400 Hz and rises. F3's that low can only be rhotic.

Barred I
[ɨ], IPA 317
On the other hand, this short vowel is too short to worry about. To the degree that it's not just 'more' /r/, I had to transcribe it as something, and following the F2-closer-to-F3 rule from Keating et al (1994), I chosed barred-i, although with an F3 that low it can't help be close to the F2. I didn't use a rhoticity hook on it, since that mostly implies a following /r/ (although there's no reason why it should). Moving on.

Lower-case K + Right Superscript H
[kʰ], IPA 109 + 404
So we have a short, but fairly solid, gap between 500 and 600 msec, with a release that heavy in the low frequencies. Very suspicious. Usually indicative of a labial release. But the F2 transition in the following vowel may be throwing off that judg(e)ment. As is the estimation of the transitions. If you look just at th transitions into the following voicing, they look like they're rising, which again suggest bilabial. But that's almost 100 msec after the release, so anything could be happening in the VOT. So the only clue that this is anything except a bilabial is the relative mushiness of the release--not nice and sharp like a typical alveolar. Labials are usually fairly sharp. I'm choosing to believe that the little blip just below 4000 Hz just after the main bit of release noise around 3750 Hz is evidence of a double burst. Of course, if I knew this was bilabial, I'd choose to ignore it. This is not an exercise in the scientific method so much as hindsight being 20/20. Don't confuse the two. Please. I couldn't live with myself as a scientist.

Lower Case W
[w], IPA 170
The thing about a [w] is that it's all transition. The noise that is visilbe in the low frequencies during the VOT is supported by F1 and F2, which you can see rise sharply once the voicing kicks in. Note the 'straight' F2 transition, typical of English (onset) [w]. Also note the F3 starts a little low, but by no means as low as an [r].

Lower-case A + Small Capital I
[aɪ], IPA 304 + 319
So abstracting away from the [w], there's nor eason for the F1 to rise to 750 or 800 Hz unless it's got some kind of target independent of the relatively flat F1 in the midrange later. So there has to be something lowish here, something compatible with the low F2 (or something odd would be happening to the F2 as well. So we're looking for something lowish and back. Then the F1 drops, so something is moving slightly higher, but critically the F2 is zooming to a maximum of aboutg 1800 Hz. Again, there's no reason of that to happen unless it's heading somewhere specific. It may or not make it, since the F3 starts coming down and knocks it out of the way, basically. So what we have here is a sequence of a lowish, backish vowel followed by a very front vowel followed by something with a low F3. But I'm getting ahead of myself.

Turned R + Syllabicity Mark
[ɹ̩], IPA 151 + 431
So I've mentioned the F3. In my head, this is syllabic, but I don't know why. I used to think I always had syllabic approximants following 'falling' diphthongs, as in 'file' and 'fire'. But 'hire', 'higher' and critically 'choir' "feel" monosyllabic to me. "Require" "re-choir". Dunno now. Hmm.

Lower-case K + Right Superscript H
[kʰ], IPA 109 + 404
Now this is a [k]. The pinchiness is a little off, since it may be that the F2 and F3 seem to be rising out of the /r/. But there's a nice double burst, the burst noise is centered in the fronted F2/F3 region. Dunno what's going on in the low frequencies. I think I'm just too close to the microphone, since something weird is happening to my input these days. The noise here looks a little sibilant, but the burst noise is just wrong for an alveolar.

Epsilon
[ɛ], IPA 303
Well, the F3 is definitely low, but in my head there's a separate vowel here. Maybe these should be transcribed as diphthongs. But anyway, the F1 is middish, the F2 is very slightly front. ANd there's an /r/ coming up, so there's a neutralization here anyway.

Turned R
[ɹ], IPA 151
Well, there it is.

Lower-case F
[f], IPA 128
Very slight frication, fairly broad band, unshaped by any resonances, and strongest, at least at the beginning, in the very low frequencies. No way this can be sibilant. The F2/F3 transitions are consistent with a labial, but that could just be the [r]. So if this turned out to be an interdental, I wouldn't be particularly surprised, although phonotactically it would be odd.

Tilde L (Dark L) + Syllabicity Mark
[ɫ̩], IPA 209 + 431
Don't ask me. But this is pretty classic. The F1 is nondescriptly mid. F2 is about as low as it could be, F3 is distinctly raised, relative to where it usually is. Lateral, and dark (velarized--the low F2). And since it's got less sonorous items on each side, must be syllabic.

Lower-case M
[m], IPA 114
Nice voicing bar, but F1 is either 'gone' or so low it's in the voicign bar. Zero below 1000 Hz, nice little pole about 1000 Hz, more zero, then another pole in the neutral F3 range. Look familiar? It should. Flat, resonant, with zeroes, must be a nasal. The 1000 Hz pole is pretty classically bilabial, and the F2 transition in the following vowel can't really be anything but bilabial also. So this one is pretty clear.

Ash
[�E6], IPA 325
So here we have another mid-to-low vowel (highish F1), a frontish but not comepletely convincingly front F2 moving, if anyything, toward neutral. Really can only be [E] or ash.

Lower-case N
[n], IPA 116
Short enough to be a nasal flap, I guess, the nasal here explaines the fuzziness of the F1 in the preceding vowel. Note th zero, tand the pole. Note also that even though the pole 'looks' like it's in the bilabial ergion, there's just a trace of something right at 1500 Hz! Woo hoo, because that's the only thing about this that makes it look alveolar. That and the flappiness, but I've been known to produce very flappy bilabials (no comments from the peanut gallery, please).

Schwa
[ə], IPA 322
Short vowel. Don't want to belabor it. Note the offset frequency of F2, near the 'locus' for alveolar transitions.

Lower-case T
[t], IPA 103
Which suggests that this gap is alvoelar. Or close.

Yogh
[ʒ], IPA 135
Broadish band of voiceless noise, sharp energy drop off below F2 and concnetrated in the F3 region. Too low to be [s] noise, so must be postalveolar.

Lower-case M
[m], IPA 114
Well, turns out there must be something here, or there's no real reason for the F2 to 'dip' into the silence the way it does (pointing down in the fricative and up in the vowel). I mean, if there were just the fricative, or just aspiration, there'd be no reason for F2 to do anything except transition. So there's something here, perhaps weakly voiced. Could be an approximant, but then it would have to be /r/, since it looks like F3 is low. But knowing wha tI do, I'll choose to ignore that... Hindsight. So the weakness might be nasality, in which case, the F2 transitions look decidedly labial. But this is hindsight too. Sorry.

Schwa
[ə], IPA 322
Vowel. Lengthend in a final syllable, but weak, very low pitched, and not really 'long'. So final lengthening notwithstanding, unstressed and reduced.

Lower-case N
[n], IPA 116
But this last syllable, if reduced, is just too long to be just a vowel and a stop. So there must be something here. Something weak. And potentially devoiced. But I have no idea how I'd tell what it is since there's not a lot of information available.

Lower-case T
[t], IPA 103
So there's something that looks like a double burst in the low frequencies (just at 2100 msec), but there's nothing else until you get up to sibilant frequencies. So on the balance this is probably a weak alveolar burst.

So putting it all together, you get "most require careful man()ch()t, where the ()s indicate some kind of vowel/syllable affair. Should be too rough to come up with something plausible, and fill in the features later.