Episode 6.3: Binary graphics and spectrographs
Added 2020-03-24 09:01:00 +0000 UTCBibliography
Japanese Phonetics Index Page
Good evening, Patrons!
This lesson breaks down the differences between binary pitch-accent guides and how they can differ from actual pronunciation by using computer generated spectrographs to look at pitch-accent patterns in spoken Japanese, as well as introducing the "subtle drop in pitch" and the effect it has on words and longer phrases.
As always, if you have any questions or concerns with regard to this lesson, please don't hesitate to leave a comment below. Thank you very much for your continued support, and good luck with your Japanese studies!
Best from Beppu,
Dōgen
Comments
wow the spectrographs are super helpful, this lesson cleared up my previous question from the earlier video as well. Thank you guys!
DoubleDoubleYea
2024-07-19 05:38:17 +0000 UTCThanks for this! I noticed the drop during the previous lesson and was going to ask you about it.
Michael
2024-06-15 15:43:53 +0000 UTCHi, Drew. JP here, Dōgen's business partner. I'm helping manage the messages here on Patreon. I can't find anything in English talking about the notation styles, but there are pages in Japanese that list and name different notation styles. One paper (available here: https://atomi.repo.nii.ac.jp/records/1417) gives 音階式 for the circles-with-lines style. The paper and this site (https://www.akenotsuki.com/kyookotoba/accent/iroiro.html) both give 線式 for the second notation style, and 丸式 for the third style shown at 0:55. One other page (https://forgetmenots.doorblog.jp/archives/55259515.html) that gives a few notations not cited in the paper or the above page. The akenotsuki.com page in particular has a pretty easy to read table with examples and a brief overview of each style, including characteristics of each. You can find more resources if you search for アクセントの表記法 or 高低アクセント書き方. Hope this helps!
Dogen
2023-10-03 02:00:21 +0000 UTCHi Dogen, do you know if there are names to distinguish the different notation styles shown at 0:55? Particularly the first two, which are the most common I see online on English resources.
Drew Edwards
2023-10-02 19:03:59 +0000 UTCI always thought this subtle drop in pitch was really interesting because it was often what stuck out to me when I was trying to learn pitch accent (different resources). I often heard it as "a drop" and had a hard time hearing the much larger drop that is used for the accent. Now I like to think of there being drops between all mora but it is the BIG drop that represents accent, which for some reason is the one I have a harder time hearing.
Joshua Ulman
2023-04-22 18:02:52 +0000 UTCHi, Crab. JP here, Dōgen's business partner. I'm helping manage the messages here on Patreon. The slower recordings, like you linked at 10:50, are used more to illustrate the individual sounds within a word or phrase. Because they're recorded much slower with sounds elongated they might sound a bit off sometimes.
Dogen
2022-12-21 00:54:22 +0000 UTCAt 10:50, is that the correct length for the first ん sound? It seemed long to me.
CRAB
2022-05-28 14:47:27 +0000 UTCGood point about being musical—this also came up a little bit in the interview I did recently with Ananya (over on YouTube). It's also good to know about the "go flat" concept in music. Thanks for mentioning this!
Dogen
2021-11-08 01:38:34 +0000 UTCNice 改善 on this ep! I liked the terracing idea from ep 29 a lot, but I've definitely found that clinging too tightly to terracing can result in a kind of robotic character (testing it out on my host mom, a native-speaker teacher). Incidentally, I imagine most of the people who "have a very good ear" are musical - I struggled a lot with 平板 pronunciation when I first started watching your vids last year because the drop in pitch over the course of the word was so obvious! But then I realized we actually have a music word for what 平板 words do: they "go flat", meaning losing the correct pitch just slightly to the lower side. Since I've been telling myself "intentionally go flat" as I shadow, 平板 words have been much easier to imitate.
Clare Kappenman
2021-10-29 14:14:23 +0000 UTCVery glad to be of service, and best of luck with your future studies!
Dogen
2021-06-07 11:53:54 +0000 UTCThanks for the reply! It's good to know that this isn't something my brain is tricking me into hearing, but something that is common enough to be noted by others. I think, most likely, I just need to trust my ears as to what I'm hearing and keep practicing. Best regards!
fred drueck
2021-06-07 01:22:15 +0000 UTCHi Fred, these are quite advanced and specific questions, so forgive me for not being able to answer in more detail. 1. Yes, there is typically a slight rise at the very beginning of 頭高 words (Vance, 2008: 153). I imagine what you said about these words needing to start with a voiced consonant in order for this to occur 100% of the time. 2. This is very difficult to say. To quote Vance from the same source as above, though the previous page: "Since many different factors influence the actual fundamental frequency of a speaker's voice, we can't realistically expect the pitch tracks of individual tokens (actual spectrographs) to match out basic intonations contours (pitch-accent graphics) much better than in Figure 7-15." Here, he's basically saying that actual spectrographs will often differ slightly from simple (typically binary) pitch-accent graphics because there are so many factors involved. It's hard to say exactly (precisely) where a downstep occurs, but I've certainly observed what you mentioned with 何が. One thing that has certainly helped me is trying to shadow / analyze your own recordings and to compare those with that of a native speaker. I've found that in actually trying to reproduce the same word / phrase, things often sort of 'click', so if you haven't tried this already then it's worth a go I imagine, at least for getting used to hearing the downstep. It also may help for you to simply think of what I refer to as the 'accent' as the 'last high mora' of the word, rather than the location of the downstep. I'm a bit loose with the terms 'accent' and 'downstep' throughout my series, as for most people breaking things down too much often ends of backfiring (trying to explain exactly where a downstep occurs often confuses people more than simply saying 'な is high and に is low', as long as the person listening also has a native recording to listen to for supplementary information). Thus, again in your case it might help to think of what I sometimes call the 'accen't or 'downstep' as the last high mora of the word, rather than the actual dramatic drop in pitch. Perhaps try thinking about things this way for a bit, then get back to me again after doing some listening with this mindset. Incidentally, Dr. Vance also talks about these concepts in detail in his book, The Sounds of Japanese, which I quoted above, so that could be worth checking out as well! Finally, I'd like to clarify that you said: "I occasionally misidentify a 尾高 word as 中高 with the accent on the second mora of the word. One such instance is the "何が” example." 何 (なに) is an 頭高 word, rather than 小高 / 中高, so I think you may have mistyped here. Hope this helps. Cheers!
Dogen
2021-06-06 12:10:36 +0000 UTCHello Dogen, First off, thank you for putting your phonetics series together, it's a great resource for trying to acquire more native-like Japanese speaking habits (as well as being able to use pitch-accent to differentiate between homonyms). Amongst many difficulties I have, one problem is that I occasionally misidentify a 尾高 (edit: should've been 頭高 as Dogen pointed out in reply) word as 中高 with the accent on the second mora of the word. One such instance is the "何が” example. Even looking at the back of the card, I still heard a slight upstep in pitch at the beginning of the word, before the pitch falls. Now, I tried generating a pitch spectogram in Audacity. to see what's going on. Assuming I've done this correctly, since this is quite likely the first time I've tried to do this, the spectogram shows I'm hearing the sounds correctly: there's a rise in pitch from the first "n" through the "a" that falls quickly after the next "n" sound and is already quite low by the "i" sound. Hopefully that is a sufficient explanation. If I could, I'd paste a picture in here, but that doesn't seem to work. Anyways, I have a theory that when a 頭高 word begins with a voiced consonant, that voiced consonant may start low and the pitch peak will occur at the following vowel. This isn't the only word where I feel like I hear this sort of thing happening, and I feel like it's likely that every time I hear it it's a similar phenomenon; the peak pitch of the word is achieved at the "trailing vowel" of the first mora and then drops immediately afterwards. However, through the initial "n" sound of the second mora, the pitch is still sliding downwards significantly. At the "i" sound the pitch is near the low level and doesn't drop much further for the final mora (the が particle). This makes things hard for me, because so far, what I've been listening for is where that drop occurs. That slide downward is the most identifiable way for me to hear the accent for most words but for this example, the drop-off occurs at the mora after the accent, not during the accent. (I also have a lot of trouble with words where the last high-pitched mora is followed by a voiceless consonant or words where there seems to be a significant trail-off for more than a single mora. In those cases, I often get the general-pattern correct but misidentify the accent point. I guess I should get around to asking a question in here at some point: First off, am I hearing the phenomenon I described with 頭高 words which start with voiced consonants correctly? Do they often start with a very quick rise in pitch and achieve peak pitch in the "vowel part" of the mora? Second off: do you have any tips on identify which mora is the accented mora in non-平板 words? Does the peak pitch always occur in the "vowel part" of the accented mora or does it sometimes occur in the "consonant part" of the accent mora? Sorry for writing this essay in the comments. Best regards!
fred drueck
2021-06-06 08:09:49 +0000 UTCHi Switz! Unfortunately I think the only way to do this is through a lot of practice, but it may be useful to try and use the previous lesson in which I use many made up words to test listeners' pitch-accent awareness. Do you mind perhaps trying to watch the previous lesson again now that you're aware of this problem to see if it helps at all? Cheers!
Dogen
2021-01-16 01:24:49 +0000 UTCGreetings Dogen! I was one of those people you mentioned who misheard some of the Heiban pattern words as Odaka pattern words due to the slight drop during the pronounciation of が -- is there any way I can get over mishearing this?
switz mka
2021-01-15 01:18:58 +0000 UTCVery glad to hear that the video helped Laura! Thank you for your support, and good luck with your Japanese classes!
Dogen
2020-09-21 05:10:16 +0000 UTCHi Dogen, thank you so much for this series! I'm so happy to have found it because pitch isn't being addressed at all in my Japanese classes (I'm getting a certificate in Japanese). This video specifically explained why I've been mishearing some of the pitches from your previous lesson. Now it makes sense.
Laura
2020-09-17 21:51:11 +0000 UTCMy pleasure Alexander! Glad to hear that you're finding the series so useful, and good luck with your future studies!
Dogen
2020-06-03 05:41:59 +0000 UTCThank you, Dogen. This made some stuff click with me. I've been studying foreign languages for a while, and it's always been frustrating how little attention was paid to pronunciation, accents, and phrasing. Figuring this stuff out while studying English has been super-frustrating. I'm extremely grateful to you and a few educators that try to help us sound not-too-horrible while speaking Japanese.
Alexander Kirko
2020-06-02 14:05:39 +0000 UTCYep!
Dogen
2020-05-05 04:49:42 +0000 UTCOh! I looked at Praat but I thought it looked more advanced than other programs. That's good to know, I'll will go try it! Thank you so much, Cheers!
Carson Render
2020-05-05 03:17:44 +0000 UTCHi Carson! I use Praat, and find that it's quite easy to use. Here'a link: http://www.fon.hum.uva.nl/praat/ Cheers!
Dogen
2020-05-05 03:13:24 +0000 UTCWhat software did you use to get the spectrographs of the native recordings and the recordings of your own speech? I want to try analyzing my own speech and other recordings of native speech with a spectrograph, because often I can't seem to tell whether or not I actually raised or lowered my pitch accurately. I've tried a couple of programs, but so far they haven't been ideal and don't seem to look as nice as the one in the video.
Carson Render
2020-05-04 20:53:21 +0000 UTCMy pleasure! Good luck with your future studies!
Dogen
2020-04-27 01:07:33 +0000 UTCThis is great - exactly what I was missing when watching the first episodes two months ago. I have a musical background and when trying to precisely mimic spoken phrases in this course, my head often subconsciously converts them into melodies made from whole & half steps and I definitely noticed that the pitch often drops gradually, but could never lay my finger on how, or rather on which scale (word vs phrase vs sentence) it happens. Thank you!
Woolookologie
2020-04-26 16:11:33 +0000 UTCHi Arthur! This wasn't on my immediate radar, and I haven't researched this in detail before, but I'll try my best to cover it in a future lesson. That said, I feel like there aren't too many major exceptions outside of すごく, though again I'll need to do more research before speaking with conviction. Thanks for the suggestion!
Dogen
2020-03-26 12:14:27 +0000 UTCHey Dogen, this isn't particularly relevant to this video, but since you're revisiting old episodes and adding additional information, I was wondering whether you were planning on making an episode that covers exceptions for i-adjectives whose く-conjugated forms don't shift backwards one mora. The most obvious one I can think of is 凄く, which has its own separate dictionary entry with a 中高 accent, but I'm positive I've heard other examples in native speech. I also notice that in the NHK accent dictionary, some examples of very common adjective conjugations list both the regular and irregular accent transformations. I would love to learn more about this, as I find I'm often unsure how to pronounce very common adjectives in certain situations. Thanks!
greyface
2020-03-25 19:20:52 +0000 UTCIndeed! haha
Dogen
2020-03-25 11:43:59 +0000 UTCBINGO!!
skabahk
2020-03-25 10:03:51 +0000 UTCHi Liam! Very glad to hear that this video cleared up all of the confusion that you had! Hope that the next video is equally helpful! Cheers!
Dogen
2020-03-25 03:10:14 +0000 UTCHi Kelvin! This is a great question—I'll try to cover this in the next lesson! Cheers!
Dogen
2020-03-25 03:09:12 +0000 UTCYou're very welcome Nathan! Glad to be of service!
Dogen
2020-03-25 03:08:22 +0000 UTCThis video cleared up So much for me! Explains excatly the areas that were confusing me. SO helpful, can't thank you enough for putting in the effort to do this. At 5:57 the particle has a drop that is greater then the downwards slope, even though it's slight..it's enough for me to think it' may be Odaka but now i know this is normal!. And at 6:50 the ま pitch bend which i thought was in anticipation of the downstep of the mora き is actually the downstep in itself. The continuous drop in pitch is normal! I have a really good understanding of to what to focus on now in order to stop the mistakes i was making. Looking foward to the next video! :)
Liam Cronin
2020-03-24 15:36:49 +0000 UTCI feel like devoicing is never explained in enough detail. Can you use spectrograms to exhibit what a devoiced downstep sounds like in comparison to a voiced downstep? I'm surprised more of your viewers do not seem to have a problem hearing a difference between the atamadaka and heiban pronunciations of 記者,
Kelvin T Nguyen
2020-03-24 15:19:50 +0000 UTCThank you so much for this lesson, Dogen! It really helped to clarify where some of the things I've been noticing actually come from. I am super excited for the next lesson as well!
Nathan Vegdahl
2020-03-24 13:37:51 +0000 UTC
