This was a very interesting exercise for me as it made more real all the very theoretical stuff that was in the reading. I have a hard time figuring out what it means that a vowel is closed, or open, or back in the mouth, or in the front. The "whispering" and "creaky voice" techniques worked to some extent to help me figure that out, but more important was making my own vowel chart.
I had some trouble with Praat and figuring out with frequencies to use (where in the striation band), but I generally picked an area in the middle and I hope that was accurate. It may not have been though, as seen in my results. I was wondering what the blue line that shows up in Praat is, with blue dots connected... it seemed like the upper border of my first formant, but I wasn't sure.
In general, my "i" is far forward, with an F2 frequency of about 2100Hz, but my "ɪ" was relatively far forward (around 2000Hz) but quite a bit more open (F1 of 430Hz versus an F1 of 320Hz for my "i"). I thought the most interesting thing was that I have very little distinction between "ɔ" and "ɑ"... the F1 frequencies for these were nearly identical, meaning that I guess my "ɔ" has migrated downwards to become a lower vowel, though it is actually farther back (slightly) than my "ɑ." I was wondering whether other people saw this shift as well?
Also, I played the McGurk Effect example for a couple of my friends and they were all quite surprised by the extent to which the visual cues dictate what we think we hear. One of my friends particularly prides herself on having a good ear, as a musician, and she was thrown off by the clip. I think I realize that this comes into effect for me particularly when speaking Spanish, which is a second language for me: I have a much easier time understanding someone if I can watch them speak Spanish than if I just hear it (for me, comprehension from reading is the easiest). I guess this might have something to do with the McGurk Effect, as I maybe map what I see the person saying onto what I know of Spanish.
Monday, October 27, 2008
Tuesday, October 14, 2008
Response to Class, October 13
I thought our discussions in class yesterday were really interesting.
I am still not sure I understand Praat as well as I could, but I'm sure if I play around with it more (and as we go further in the quarter with some of the more acoustic stuff) I will start to get it. It was really interesting to compare notes on how that last assignment turned out. I thought it funny that so many people had tried for a Southern accent (including me); I guess this was because the accent is so strongly "regionally marked" and that it's not as hard to imitate, even if ithe imitation is horrible. Also, I thought the cross-cultural stuff we discussed was fascinating: Nikolas' comment about the Brazilian equivalent of "like" that he puts into English conversation, Fatima's comment about the Arabic "you know," and then the discussion on the German "isn't it?" that's sort of added on to things, along with Meghan's husband's semantical switch. I recently watched a clip of Russell Peters, a Canadian comedian, on a similar topic with the Chinese equivalent of the word "like," which is something like "nega." I am also looking forward to hearing more about some of the projects which I think will come out really well; if anyone has any suggestions on mine, which is to see how well people can determine other people's accents, please let me know.
Speaking of accents, I got 8 out of the 10 different accent types correct in the quiz. The ones I missed were #5 and #8, which were a Hispanic and a black person, respectively. The last one was very obviously Indian, with very few aspirated consonants and really closed "o"s (like in most Indian languages), along with a difference in the structure when he said "fleece as white as the snow." The black Caribbean person also had a pretty distinct accent, with the "a" in Mary and the third "e" in everywhere pronounced similarly (almost as I'd say the "ai" in fairy), along with an "o" that sort of became a "u", as in "go'u" and "sno'u." Lastly, his "r" was a bit different: it was like the "r" in butter if you were enunciating it properly. I know that if I knew more Caribbeans (or, better yet, if I were Caribbean) I'd be able to tell where in the Caribbean he's from, as there are subtle distinctions between Jamaicans, Trinidadians, Barbadians, Belizeans, etc. The African American people were relatively easy to identify. There was something especially pertaining to the "a" sound, which the first woman moved to almost an "iaa" ("lyamb") and the second man stretched out too. The "o" had a clear diphthong as well, which the Hispanic people, for example, did not have as much of. In fact, I think I determined the Hispanic people in part based on the way they said their "o"s at the end of the phrases, which was an "o" sound without as much "u". The Arab man was also relatively easy for me to pick out, perhaps because I am slightly more familiar with the accent; again, the "a" in Mary and the third "e" in everywhere had a different intonation from other American English heard in the trial, though it was not the same as the Caribbean -- more of a straight "e" as in the IPA.
This was an interesting exercise because while I think they chose people with relatively strong accents from their respective groups, I didn't get all of them and I know that I do make judgments about someone's ethnicity from their accent. This, and it relates to the matched guise experiment we watched in the housing PSA, is important in realizing that we can't make those kinds of judgments accurately always, and that some people out there do take it to the next level where they discriminate based on the assumptions they make.
I am still not sure I understand Praat as well as I could, but I'm sure if I play around with it more (and as we go further in the quarter with some of the more acoustic stuff) I will start to get it. It was really interesting to compare notes on how that last assignment turned out. I thought it funny that so many people had tried for a Southern accent (including me); I guess this was because the accent is so strongly "regionally marked" and that it's not as hard to imitate, even if ithe imitation is horrible. Also, I thought the cross-cultural stuff we discussed was fascinating: Nikolas' comment about the Brazilian equivalent of "like" that he puts into English conversation, Fatima's comment about the Arabic "you know," and then the discussion on the German "isn't it?" that's sort of added on to things, along with Meghan's husband's semantical switch. I recently watched a clip of Russell Peters, a Canadian comedian, on a similar topic with the Chinese equivalent of the word "like," which is something like "nega." I am also looking forward to hearing more about some of the projects which I think will come out really well; if anyone has any suggestions on mine, which is to see how well people can determine other people's accents, please let me know.
Speaking of accents, I got 8 out of the 10 different accent types correct in the quiz. The ones I missed were #5 and #8, which were a Hispanic and a black person, respectively. The last one was very obviously Indian, with very few aspirated consonants and really closed "o"s (like in most Indian languages), along with a difference in the structure when he said "fleece as white as the snow." The black Caribbean person also had a pretty distinct accent, with the "a" in Mary and the third "e" in everywhere pronounced similarly (almost as I'd say the "ai" in fairy), along with an "o" that sort of became a "u", as in "go'u" and "sno'u." Lastly, his "r" was a bit different: it was like the "r" in butter if you were enunciating it properly. I know that if I knew more Caribbeans (or, better yet, if I were Caribbean) I'd be able to tell where in the Caribbean he's from, as there are subtle distinctions between Jamaicans, Trinidadians, Barbadians, Belizeans, etc. The African American people were relatively easy to identify. There was something especially pertaining to the "a" sound, which the first woman moved to almost an "iaa" ("lyamb") and the second man stretched out too. The "o" had a clear diphthong as well, which the Hispanic people, for example, did not have as much of. In fact, I think I determined the Hispanic people in part based on the way they said their "o"s at the end of the phrases, which was an "o" sound without as much "u". The Arab man was also relatively easy for me to pick out, perhaps because I am slightly more familiar with the accent; again, the "a" in Mary and the third "e" in everywhere had a different intonation from other American English heard in the trial, though it was not the same as the Caribbean -- more of a straight "e" as in the IPA.
This was an interesting exercise because while I think they chose people with relatively strong accents from their respective groups, I didn't get all of them and I know that I do make judgments about someone's ethnicity from their accent. This, and it relates to the matched guise experiment we watched in the housing PSA, is important in realizing that we can't make those kinds of judgments accurately always, and that some people out there do take it to the next level where they discriminate based on the assumptions they make.
Saturday, October 11, 2008
Praat Comparison
This is my blog regarding the homework assigned from last class, the comparison of a simple sentence in two different pronunciations using Praat. The sentence, "The butter spilled on the cot," is a simple one that has a variety of vowel sounds in it. The schwa occurs in "butter" -- which also has the glottal "tt" -- and in "the." There's a short "i" sound in spilled and the "o" in cot and in "on" is often pronounced in a variety of ways across different dialects.
I attempted to use a Southern dialect to say the sentence. I used some of what we've read and some of what we discussed in class about vowel shifts to affect my accent. The short "i" in spilled I moved up towards the front and top of my mouth to produce more of an "iy" sound. This is visible in the spectrograph clearly. The pattern as compared to the "i" in my "normal" accent is much darker, and there is a downward (convex) arch along with two dark horizontal bands above and one below. In my normal accent, this curve exists but it is more concave and blends in with the bottom horizontal band (probably because my natural "i" is not moved forward).
I think the "o" is represented by the amount the final band is curved because in the "Southern" version there is more of a curve and I can distinctly hear a more stretched out, broader "o" sound in cot than in my normal rendition. I also notice that the "c" sound I made in the Southern accent is more drawn out and aspirated, which is seen in the spectrograph a stop in the spectography whereas in my normal one the graph is more blended in the word "cot."
Also, in the "Southern accent" the words are distinct but the sounds seem to blur together (ie. the bars aren't as distinct), which is probably the effect of the "drawl" that we hear. Related to this is the fact that it took longer to say the same sentence with the accent, (by 0.7s) something I did not do intentionally.
The word "butter" is quite different between the two versions. In the "normal" the amplitude is relatively high for the "bu" and then lower for "tter", while in the Southern it's all quite the same. Also, the drawing in the normal version shows three distinct areas of darkness, one sloping up during the "b", then a hazy middle area for the glottal "tt" and then a higher dark area for "er." In the Southern one these three areas are the same, but they are actually connected. The middle area is also much less hazy and clear as a dark spot. I think this might be from the fact that it's less glottal and approaches a "d" sound almost, which connects with the "er."
The last thing that I thought was interesting (which was the same in both) was the way I grouped words. I can tell where "The" starts and ends in each, but then "butter sp--" is relatively contiguous -- though you can make out which is which. There is then a space and then "-illed on the" and then, separately (probably due to the aspirated "c"), "cot." I would have thought that all my words would be relatively distinct or relatively together, not that some would be together and others apart. I guess this might vary from person to person though.
I attempted to use a Southern dialect to say the sentence. I used some of what we've read and some of what we discussed in class about vowel shifts to affect my accent. The short "i" in spilled I moved up towards the front and top of my mouth to produce more of an "iy" sound. This is visible in the spectrograph clearly. The pattern as compared to the "i" in my "normal" accent is much darker, and there is a downward (convex) arch along with two dark horizontal bands above and one below. In my normal accent, this curve exists but it is more concave and blends in with the bottom horizontal band (probably because my natural "i" is not moved forward).
I think the "o" is represented by the amount the final band is curved because in the "Southern" version there is more of a curve and I can distinctly hear a more stretched out, broader "o" sound in cot than in my normal rendition. I also notice that the "c" sound I made in the Southern accent is more drawn out and aspirated, which is seen in the spectrograph a stop in the spectography whereas in my normal one the graph is more blended in the word "cot."
Also, in the "Southern accent" the words are distinct but the sounds seem to blur together (ie. the bars aren't as distinct), which is probably the effect of the "drawl" that we hear. Related to this is the fact that it took longer to say the same sentence with the accent, (by 0.7s) something I did not do intentionally.
The word "butter" is quite different between the two versions. In the "normal" the amplitude is relatively high for the "bu" and then lower for "tter", while in the Southern it's all quite the same. Also, the drawing in the normal version shows three distinct areas of darkness, one sloping up during the "b", then a hazy middle area for the glottal "tt" and then a higher dark area for "er." In the Southern one these three areas are the same, but they are actually connected. The middle area is also much less hazy and clear as a dark spot. I think this might be from the fact that it's less glottal and approaches a "d" sound almost, which connects with the "er."
The last thing that I thought was interesting (which was the same in both) was the way I grouped words. I can tell where "The" starts and ends in each, but then "butter sp--" is relatively contiguous -- though you can make out which is which. There is then a space and then "-illed on the" and then, separately (probably due to the aspirated "c"), "cot." I would have thought that all my words would be relatively distinct or relatively together, not that some would be together and others apart. I guess this might vary from person to person though.
Subscribe to:
Comments (Atom)