Solution for (mid-)July 2005

labelled spectrogram
"They don't know where to go."

Eth
[ð], IPA 131
I worked really hard at making an unstopped, fully voiced initial Eth, and I'm still not 100% successful. But look at that voicing. More than 100 msec of it. But the visible frication is, well, as weak as it should be, but it really only creep sin in the last 25 msec or so, and really only once the voicing in the upper formants starts to creep in which might as well be vowel. So I don't know what to do. We've got voicing. We've got fricative, sort of. We've got transitions into the following vowel consistent with alveolar (less so than velar or bilabial--the F3 is just sort of sitting there, where for either velars or bilabials it should start lower), but there's no sense in which the fricative looks sibilant. So we're talking some kind of coronal fricative, nonsibilant. And voiced. Narrows it down quite a bit, actually.

Lower-Case E + Small Capital I
[eɪ], IPA 302 + 319
Well, I'm not sure there's a height movement happening (that F1 looks pretty flat to me) but the F2 definitely starts one place and ends up another. I'm not sure how much of that is transition from the consonant, and how much is diphthong, but the fact that it's straight and not curved suggests two targets rather than one target with a longish transition toward it. But I don't know. Mid vowel, judging from the F1, front, judging from the F2, and if it's a diphthong, it has to be tense.

Lower-Case D
[c], IPA 104
Looking at this again, I'm wondering if I mis-transcribed it. We've got a shortish gap, with pretty full voicing throughout. There's even some evidence of resonance through the gap, so maybe I should have transcribed as a flap. But I didn't, so there you go. It's a gap, so if we take it as a plosive, then we have to decide which one. Well, voiced. No velar pinch, which leaves bilabial or alveolar. ANd the transitions don't really look bilabial.

Lower-Case O + Upsilon
[nʊ], IPA 307 + 321
Ah, movement. This vowel, and the upcoming couple of consonants, were the whole point of this utterance. Okay, F1 is a little higher in frequency than the previous vowel, but it's still in the mid range. The F2 starts up around 1750 Hz but then falls sharply to a low just about 1000 Hz. So this is mid. It seems to start front, but falls outlandishly sharply towards definitely back and round. I don't know how much of that is allophonic and how much is just me and my screwed up back vowels. Something to think about as ew move on. But anyway, the only things that can end up that back in my dialect of English are phonemically/historically back. And round. So this is an /o/, an id it looks like a diphthong in my speech. Remember that. For the record, I think I transcribed it as nasalized (at least the offglide before I decided that there was a nasal consonant in the following coda, distinct from the onset after that. More below.

Lower-Case N + Glottal Stop
[nʔ], IPA 116 + 113
At about 450 msec, the spectrum changes abruptly. All the the energy between the formants drops out. The movemnet in F2 and F3 changes directon or slope or both pretty suddenly. But there's 50 msec or so of solid voicing, with resonance (let's say 'pole') at about 1000 Hz, something going on at 2500 Hz or thereabouts, even energy at 3500 Hz. Okay, so The absence of energy is evidence of a zero of somekind coming in, which is evidence of nasality. So far so good. I transcribed as a nasal with a glottal stop, knowing what the utterance was, but actually, I have no good way of knowing it's an alveolar nasal. This could be a nasalized [ʊ]. That would explain the low pole (it would be F2). The evidence for glottality coems from the ragged striations in F3/F4 up around 2500, I guess the real clue that tehre's something going on here, coda-wise, is the abrupt change at about 480msec. More below.

Lower-Case N
[nʔ], IPA 116 + 113
Now this is a nasal. I'ts fully voiced, the zero(es) have expanded to kill even the lower pole. And the resonances are flat. Frankly, it still looks like a bilabial, but I'm *really* hoping that that's just coarticulatory rounding. The point of this part of the spectogram for me was to see what this transition looked like. When you've got one continuous closure (I suppose) at the alveolar ridge, the velum down continuously. I'm sure there's tongue and jaw gestures happening to distinguish the coda nasal from the onset one, but I'm not sure how they'll show up. I'd assumed that we'd see evidence of glottality and a change in the amplitude of voicing. Well, neither of those is amazingly robust. What I definitely see is the change in the amplitude of the zero, which I'm going to have to think about some more.

Lower-Case O + Upsilon
[nʊ], IPA 307 + 321
Well, we've got an F1 in the mid range (around 500 Hz) but this time, it's moving slightly downward. The F2 is starting slightly back of centered (i.e. below 1500 Hz), but again is moving downward. So this is starting mid and vaguely back, and moving backer (and/or rounder) and higher. Once again, this is /o/.

Lower Case W
[w], IPA 170
Well, that's the lowest F2 you're ever likely to see from me. The F1 isn't quite as low as I'd expect for an intervocalic (and presumably onset) /w/, but oh well this isn't an amazingly strong boundary, in spite of its syntax. Which we don't know about yet. Oops. Okay, so we've got an F1 in the mid to higher mid range, an F2 extremely low indicating extreme backness and/or rounding. F3 not doing much. So this is something resonant, backish and roundish. The lowered energy in the high frequencies suggests a consonant. Given that it's obviously surrounded by things which are 'more' vocalic, this is something vowel-like acting like a consonant, i.e. an approximant. So if it's a choice between /u/ (or /o/ or something like that) and /w/, we know which to pick.

Epsilon
[ɛ], IPA 303
Well, this vowel is mostly transition, but from what we can see of this syllable, the F3 is still hovering in the mid-range. Abstracting away from the preceding /w/, the F2 seems to be heading toward the front space. The F3 is dropping, eventually coming to a minimum around 850 msec, so we'll save that moment for later. But the point is the F2 is heading as high as it can until it stats to run into the lowering F3. So this is middish, and frontish. And given the following segment (that F3 thing) it's probably not worth asking whether it's tense/long/diphthongized or lax/short/centralizing.

Turned R
[ɹ], IPA 151
So the F3, which up intil about 850 msec is hovering between 2300 and 2700 Hz dives to around 2000. Think that's not low enough? Get real. Especially for a coda /r/ with a front vowel. So the reason we're not worried about tense or lax in the previous vowel is that the tense/lax distinction neutralizes before /r/. La.

Lower-Case T + Right Superscript H
[tʰ], IPA 103 + 404
So there's a couple of very widespread pulses in this gap (from about 900-950msec or so), but given the aspiration following the release, I'd not worry about it. The release occurs at about 950msec, and there's not only at least 50 msec of VOT, but there's higher frequency noise (look at that around 3500 Hz for 100 msec after that. And that noise is high frequency, broad band and *very* strong. So that's sibilant release, which means this plosive has to have been a) behind the teeth and b) close enough to produce noise upon release that resembles an [s]. Hence an aspirated /t/.

Barred I
[ɨ], IPA 317
So amid all that high frequency noise, there's a neeny little vowel. GIven how short it is, it's not worth thinking too long about. But it's mid or higher-mid judging from the F1, and slightly front, judging from the F2. But it's reduced, so treat it that way.

Lower-Case G
[g], IPA 110
The gap here starts at agbout 1050msec and goes on to, well, after 1100 msec. And the voicing is pretty strong and fairly consistent, considering it looks like a serious plosive and not something slushymushy like my plosives so often are. The burst is very broadband, and it's a little tilted t the higher frequencies. But the burst isn't 'sharp' the way the [t] release was. It could almost be 'double'. Hmm. And there's a lot of energy just about 1700 Hz or so, rather than being clearly concentrated higher up. So while I'll admit this is ambiguous, but the slightl low F3 might be evidence of velar pinch. It's unlikely to be coarticulation with the previous [r], just because there's at least a whole syllable in between.

Lower-Case O + Upsilon
[nʊ], IPA 307 + 321
And finally here's another /o/. Again it seem sto be mostly mid throughout (the F1 is basically flat at 500 Hz for 400 msec or so. It starts out sort of central (in the mid F2 range) and moves decidedly backer/rounder (lower in frequency). So I guess my /o/ is a diphthong. My /e/ is still open to interpretation....