Pronunciation of the vowel 'u'

Hey there,

Something that comes up again and again for me when shadowing native speakers is the having trouble with the ‘u’ sound. As in any of う, く, す, つ, etc.

To me, there seem to be two ways different ways speakers will pronounce the ‘u’. It’s hard to put in words (as all discussions about sound are) but one of them sounds sort of like ‘oof’. And the other … actually I can’t think of any sound in english that is similar to it. I honestly think the sound doesn’t exist in english. In germany we do have this sound (or at least almost that sound) and its ü. So that’s what I’ll refer to it for now.

An easier way to get across what I mean is audio of course. And I happen to come across a perfect video to demonstrate what I’m talking about here. Notice the difference in pronunciation of る when the whole word is said versus when each mora is said one at a time.

Another example: Here’s the word 来週 (らいしゅう) pronounced on forvo. The first four speakers seem to pronounce it as ü. Then the user skent pronounces it like u.

Has anyone else noticed this? From listening to native speech, I seem to hear ü more often than u. But then a lot of speakers do a sound that sort of in between, so it seems more like a spectrum to me. On wanikani the audio sounds mostly like ü to me. But there is native speaker I sometimes listen to on youtube and he very clearly pronounces it as u every time. Is it an accent/area thing? Is one considered more feminine/manly? (I seem to hear u mostly from men) How do you pronounce it?

Or am I taking crazy pills and there isn’t really a difference at all? :sweat_smile:


I honestly don’t hear something different in your various examples, but it seems like with basically everything there are going to be allophones. That is, there will be an accepted range of slightly different sounds between different native speakers, and often they won’t even be aware that a range exists.

But the takeaway from that is, it literally won’t matter what you choose, as long as you are mimicking natives, since that will be something close enough to fall in the range.

1 Like

That’s so crazy to me, because both examples sound so very different to me. I’ve wondered before why I can’t seem to find other people with the same problem on google. I’ve found one discussion on this on a german japanese learning forum at least, but sadly the forum did not seem to be very active.

As I explained, one of the sounds I’m trying to describe doesn’t exist in english, so perhaps an english native speakers ears aren’t adjusted to hearing the difference (similarly to how japanese people have trouble differentiation between l and r).

I played back these clips to another german friend of mine and he agreed on it being different at least :thinking:

I also hear a difference between “skent” and all the others. I’m also German btw. skent sounds like an older guy to me (I wouldn’t really call it dialect though), while the others sound like standard Japanese.
I pronounce う as a mix of ü and u, maybe leaning more towards ü.

I go with ü over u since a) it seems native speakers statistically use ü more often than u as well, especially speakers of the ‘standard’ Tokyo dialect, and b) most people I hear saying u over ü are native English speakers who speak Japanese as a second language. Since English doesn’t have the ü sound, it’s relatively hard for native English speakers to wrap their mouths around that pronunciation, compared to native speakers of German, Dutch, French and others where that sound is common.

I guess in my mind saying u instead of ü makes me sound more off-center than native Japanese, and that’s what I try to avoid in any language.

All in all, I don’t think it matters too much. You’ll probably be understood by native speakers about as well or as poorly either way. And a lot of people probably won’t even be able to tell the difference unless they pay very close attention. If a sound doesn’t exist in your native language, you’re much less sensitive to the nuances when hearing it.

1 Like

Bingo. As children, as we get adjusted to speaking/hearing our native language, our brains essentially ‘tune in’ to the things that matter, the things that are important to distinguish. And as we get older, that tuning gets locked in, and the subtle differences we never learned to pay attention to (because we didn’t need to) fade away and become genuinely hard to notice. Each language trains a child’s brain differently, so you get people who genuinely can’t tell sounds apart that are very different to other people.

The idea that many Japanese people have difficulty distinguishing L and R sounds as they exist in the languages I grew up with is bizarre to me. And similarly, most westerners have difficulties with the subtleties of Korean pronunciation, which in turn seems bizarre to many native Koreans to whom it’s so obvious.

1 Like

What would be something that doesn’t exist in German, but does in English? I guess when you think of English speakers mimicking German accents, I think the “th” sounds take the brunt of it. Is that fair to say, that the English “th” sounds don’t exist in German, or are different in some way?

The ‘th’ sound exists in German and Dutch (and probably a few others) in a degenerate form. In words like ‘thema’ or ‘theologie’ the ‘h’ is pronounced slightly aspirated. English has turned that into its own big thing, but in German and Dutch it’s so subtle that most people don’t even notice it. But it’s there. The way English speakers satirize German ‘th’ as “zee Germans” is mostly an exaggeration, but it works well enough for ‘Allo, Allo’ :wink:

The most prominent sounds that exist in English but not in German are actually vowels, or vowel contractions. We like to think there are 5 vowels, but English actually has about 20. We just write them with just the 5 letters (and sometimes y). A few vowel sounds that I don’t think show up in German:

The ‘a’ in ‘land’.
The ‘e’ in ‘news’.
The ‘o’ in ‘phone’.

And the typically English non-rolling ‘r’ is also very absent in most other languages.


The ‘th’ sound is an obvious one, but it’s distinct enough that we german’s won’t confuse ‘zat’ with ‘that’.

There is one sort of sound that we do not have, and it’s actually not well know to most germans who speak english, because it’s hard to hear for us. Basically, in german, no words end with a ‘soft’ sound like ‘d’, ‘b’ or ‘g’. All our words, even those ending with ‘d’, ‘b’ or ‘g’ have a the ‘hard’ ending ‘t’, ‘p’ and ‘k’. Some people get the these sounds right after some time unconciously, but the difference between ending a word with d or t is a mistake I hear all the time and still make myself.

I suspect this is a part of why german sounds so ‘angry’ to people. :upside_down_face:

So if you wan’t to complete your german accent say ‘dok’ instead of dog, ‘stant’ instead of stand, etc. :grin:


Sorry to be pedantic, but that’s wrong. The th in words like Thema or Theologie is an orthographic relic from Greek, but it doesn’t change the way the sound is pronounced in standard German. There is no difference in pronunciation between Thema on the one hand and Tennis or Teppich on the other, all three are pronounced with a slightly aspirated [tʰ].

English also didn’t turn th, or /θ/, into “its own big thing”, but instead is one of the few Germanic languages that retained the proto-Germanic /θ/ sound as a phoneme, as did Icelandic. All other Germanic languages lost /θ/, a process which is estimated to have taken place between the 9th and 10th century, when dental fricatives changed to plosives. Unlike other shifts that affected High German only, this particular shift extended across the entire continental Germanic language area, which can be neatly illustrated using a couple of examples:

EN thank, IS þakka vs. DE danken, NL danken, SE tacka
EN thin, IS þunnur vs. DE dünn, NL dun, SE tunn
EN south, IS suður vs. DE Süd, NL zuid, SE söder

You’ll see that the English /θ/ first and foremost occurs in words of Germanic origin as opposed to words of Greek origin like theater. Those didn’t enter Germanic languages until the 15th to 16th century (sometimes even later than that), hundreds of years after the shift from dental /θ/ to alveolar /d/ and /t/ had already taken place.

Words like theater are pronounced with a /θ/ in English as a spelling pronunciation since they primarily entered the language through written sources and th already existed as the codified spelling of the sound /θ/. Most of these loans actually entered English via French, which lacked a /θ/ sound as well. German and Dutch borrowed these words from French as well, which is how the spelling was passed on, but there’s no link to the sound /θ/.


To give another viewpoint on this, I don’t speak German, but I can hear a difference between the sounds in the second example, although not the first one. I’m guessing it’s a case of the normal う being said without very rounded lips, but in the case of the outlier it sounds to me much more rounded. Like Leebo said, I think it’s just a difference of how different speakers pronounce it.

If I am right about the more rounded lips thing, Dogen says in lesson #35 of his Phonetics series that “The lips should be completely neutral while saying this vowel. There is no lip protruding or rounding.”, so it’s probably best to stick to sounding like the 4 speakers in the second example.


This would also be a big difference between Japanese [ɯ], which is indeed an unrounded vowel, and German and French [ʏ] as in Glück or tu, which is rounded. The other differences being that [ɯ] is pronounced further back in the mouth than [ʏ] and with the tongue closer to the roof of the mouth. [ɯ] is basically the unrounded version of [u] as in German Fuß or French fou and therefore closer to [u] than [ʏ].


No, there’s a massive difference. I particularly like the 食べる video that’s actually close to using 3 different pronunciations of the う vowel. ü for the full word, u for the mora and then more similar to the Dutch uh for the repeat of the full word (uu, oe, uh in Dutch). A lot of the audio files on WK use a sound that’s closest to the ü, although it’s kind of a mixture of the three sounds all the time. I’ve tried shaping my mouth for an uh, thinking an ü and then sort of grunting an u from the throat and find that this usually gets pretty close to the desired audio. YMMV.

1 Like

And yet Japanese people will probably tell you that, no, they’re all just う. So whether there’s a difference at all and whether it’s massive is all a matter of perspective.


I guess it should make me less worried about pronouncing it - apparently as long as it’s close it should be understandable.

1 Like

An easier way to get across what I mean is audio of course. And I happen to come across a perfect video to demonstrate what I’m talking about here . Notice the difference in pronunciation of る when the whole word is said versus when each mora is said one at a time.

I do notice that too. Could this explanation on Wikipedia help? (I’m copying the explanation below:)

/u/ is a close near-back vowel with the lips unrounded ([ɯ̟])[36][37] or compressed ([ɯ̟ᵝ]).[4][38] When compressed, it is pronounced with the side portions of the lips in contact but with no salient protrusion. In conversational speech, compression may be weakened or completely dropped.


I also hear a very pronounced difference in the forvo example, with the 4 first ones being more standard. I guess normally the japanese u is somewhere between u and the finnish y (which is similar but even more to the front than ü). Like in this video: How to pronounce y - YouTube

In regards to the (un)rounded lips. I did come across that before but personally I can make both sounds using both rounded and unrounded lips. This does give both sounds a different quality, but it’s not what makes them different from each other.

It’s definitely good to know though, because the unrounded lips definitely makes it sound more japanese u.

1 Like

I listened to the 来週 link and the difference that I hear is that the ‘skent’ version vocalises しゅう a bit after the others. For the others, the vowel slide from い to う is clearer.

I didn’t get the feeling that any of them had rounded their lips.

I can’t be the only one… In the 来週 example… I don’t hear the “r” sound when Skent speaks. For me it sounds like がいしゅう。:scream:

Am I deaf or is this some weird regional thing?