Depending on which font(s) you have installed, one or the other symbol in paragraph headings may not display correctly. To properly view the phonetic symbols on this page, you must have one of the following fonts installed on your system.
Starting just before 100 msec, there's a very clear voicing bar. Too clear, actually, but whatever. It's weak, but there's a trace of resonant information above. Not really enough to indicate full, open, resonating chambers, so this is some kind of close approximant or fricative. The choice of fricative comes from the 25 msec of noise or so just before 200 msec. So this is sort of affricated, but not really. Just bear in mind this pattern as the logical output of initial fortitian of an initial Eth. It's voiced, it's vaguely fricative (perhaps mostly stopped in its first phase, but clearly fricative in its release). It looks coronal, judging from the F2 and F3 transitions (F2 'locus'ed between 1700 and 1800 Hz, F3 high), but not at all sibilant. The noise we see is broad band and not particularly high in amplitude (compared with other fricatives in the spectrogram). So coronal, but not sibilant, fricative, voiced.
Lower-case E + Small Capital I,
IPA 302 + 319,
This is pretty typical of my /e/s, a vowel that is quite high (and notice the flat F1 contour), starting quite front and just getting fronter. So if you really want this to be a mid vowel, sorry. It ain't. See Hagiwara (1997) for some interesting pictures...
Lower-case L + Mid Tilde,
IPA 155 + 428,
This isn't the darkest /l/ I've ever produced, but it definititely ain't light, clear, or plain, (whichever you pick it still ain't it). So I've marked it as velarized. The relatively low (or at the very least not at all raised) F2 is the factor in darkness, and, well, I've made my choice. The amplitude discontinuity (at about 300 msec going in and about 375 going out) suggests a degree of closure, and there's just too much energy in the lower frequencies to make a good zero (which we'd expect with a nasal). The raised F3 (if you wish really really hard, have lived an honest life, or are pure of heart) and the vaguely raised F4 (if you can make it out) is/are good cues to the /l/ over the other approximants.
Lower-case A + Small Capital I
IPA 304 + 319,
The F1 starts out quite high, 800 Hz or so, indicating a very low vowel. F2 seems to start out rather low (although on further consideration, it's pretty neutral. So we're looking for something that starts out as a low vowel of, well, non-back quality. I think that describes IPA [a] (Cardinal 4) pretty well, especially in English, where the other choice for a non-back low vowel is Ash, but hopefully would look more front. And not as low. But whatever. After about 400 msec or so the formants transition pretty clearly, F1 dropping in the last third of the vowel duration (indicating raising) and F2 raising into the very high frequencies (for F2) indicating extreme fronting. So this is a diphthong that moves from [a] to high and front.
Lower-case K + Right Superscript H,
IPA 109 + 404,
The other thing worth noting about the preceding vowel si that the F3 starts to drop towards the end, as the F2 raises. In fact the F2 is heading up to a frequency that is well into the [i] range, as far as frontness goes, but it's heading up to meet the F3, which is still slightly raised presumably from the preceding /l/ or even the off-glide before that. Anyway, F2 and F3 coming together like that is pretty classic 'velar pinch', which is mirrored (in reverse) in the shaping of the aspiration noise on this segments release (550-600 msec or so). So this must be a velar. And pretty clealry voiceless and strongly released, which I transcribe as aspiration, even though this isn't classic VOT-style aspiration.
There's about three pulsey looking things in the trailing end of the preceding plosive release. So this might be a teeny bit of voicing and I should have marked a tiny vowel here. But I didn't, because these pulsey thingies are just my vocal folds flapping around as I close them. I'm not sure why I wanted a glottal stop here, since this isn't really a very strong prosodic break, but there it is. There's also some irregularity in the voicing o the following vowel, which I interpret as further evidence of glottality here.
Lower-case A + Small Capital I
IPA 304 + 319,
Another one of these dipthongs, this one with less steady state and a much clearer F1 transition, but otherwise pretty similar.
Very broad-band noise. This segment is a little ambiguous, since it's not obvious where this noise is centered, i.e. it doesn't obvioiusly get stronger in the higher frequencies, nor does it obviously have a lot of energy in the F2-F3 region. These two are the most positive cues to either /s/ness or Eshness (I don't really feel like switching fonts in the middle of sentence, so work with me here) respectively. It doesn't help that the energy drops off suddnely below 1500 Hz, which is usually a good cue for Eshness (Esh seems to almost always have a zero in the low frequencies). So there you go. If you guessed Esh here, I can't fault you. ALthough It's worth pointing out that the energy drop off is similar in profile (if not in absolute distribution) to the following plosive release/aspiration, whichis pretty clearly /s/-like. Then again, it's not so far off from a fricative coming up later that will confuse everyone.
Lower-case T + Right Superscript H,
The closure portion of this thing lasts about 60 msec, which is pretty long for one of my stops. This may be evidence of gemination, but I'm not convinced I say 'iced tea' and not 'ice tea'. I definitely tried to say 'iced tea', but I'm not sure I did. Anyway, the F2 transitions, if there are any, are consistent with an alveolar articulation, and not obviously indicative of velar or bilabial articulations. That's my story and I'm sticking to it. The real give away is the high-frequency center of this noise, more obviouisly [s]-like than anything else.
Well, you couldn't ask for a better [i], unless it had a narrower-band F1. It's really broad-band, but I think the center of the F1 is about 300 or 350 Hz or so, which is pretty low, I guess. The F2 is freakishly high, up into the neutral range for F3. So if you don't regard this as hyperarticulated (and probably stressed), you've missed something. Not to mention how long this thing is. Egad. Anyway low F1 (high vowel), high F2 (front vowel).
Woof. I wouldn't want to be riding this F2 through this spectrogram. Check out that drop from about 1300 to 1350 msec. So the F1 stays dead flat, indicating high or close articulation that doesn't vary a whole lot. The F2 goes from absurdly high to absurdly low (relatively speaking) indicating extreme backness and rounding (probably both). The F3 doesn't reall move much out of neutral (once the F2 gets out of the way), so this can only be [w].
Well, this vowel is quite short and it's in a syllable next to something very clearly stressed. So chances are good that this vowel is unstressed and therefore reduced, and therefore doesn't carry a lot of information. So I call it schwa and move on.
Fricative. That part is clear. Broad-band and top-heavy so this looks sort of /s/ like, though it doesn't have nearly the amplitude that the other [s]s in this spectrogram have. So if you think this is an [s], I guess I won't argue with you. There is, however, this little matter of the F2 transition into it, which just doesn't look alveolar (although dependng on what you think is going on on the other side, this oculd be transitional to that). And then theres the teeny littly gap at about 1550 msec, followed by a fairly clear release burst that doesn't really look alveolar either. So we've got something that doesn't really look alveolar, but it doesn't really look like anything else. Sorry, that's the best I can do, but I could go on for a while about how an interdental could look like an alveolar that isn't an alveolar. But I won't.
Lower-case L + Mid Tilde,
IPA 155 + 428,
Eseentially, this looks the same as the last one, so there's not much point in repeating myself. This one looks more like a nasal, with its abrupt edges and such, but the zero (if that's what that is between F2 and the very raised F3) is in the wrong place. Nasals almost always have a nasal between the F1/voicing bar and the F2/nasal pole.
Well, this is about as low an Epsilon a I have ever seen. It's even lower (higher F1) than the [a] in the first diphthong. The F2 is neutral, and the F3 tells us nothing. So I supposed this looks like an Ash, but it's sort of short and flat to be an Ash. Which leaves Epsilon as the closest plausible vowel. Maybe I'm starting to do Canadian Shift/Lax Vowel Lowering.
See, now *this* is a nasal. Voicing bar or F1, about 300 Hz, Zero, Pole/F2 thing about 1200 Hz or so. The pole is a little high for my [m], but not high enough for my [n]. But the F2 and F3 (and F4) of the surrounding vowels all point downward, which is pretty good evidence of bilabial articulation.
Schwa. Not sufficiently different from the previous vowel to be worth arguing about, so I I won't.
Another nasal, thi sone is complicated by the fact that my amplitude goes down at the end of the utterance anyway. But notice the voicing bar, the zero(es) and the evidence (if you wish hard enough) of a resonance at about 1500 Hz, which is about right for my [n]. Note that wherever you think this resonance is, it's different from the preceding nasal, so if there's a choice to be made, they're different.