Correct pronunciation of voiceless bilabial fricative [ɸ]

Question

I found that there are two different pronunciations for the [ɸ] phone. For example, in Japanese compare those two pronunciations for same word 冬 (for which I believe will has [ɸɯjɯ] as phonetic transcription):

• https://jisho.org/search/%E5%86%AC. To my ear, this one sounds more like [pʰɯjɯ].
• https://translate.google.com/#view=home&op=translate&sl=ja&tl=en&text=fuyu. . To my ear, this one sounds more like [kʰɯjɯ].

In this wikipedia page, which provides the audio sample for the sound (near the top right of the page), I also found that phone is pronounced much similar to the first link in the list above.

However, some other sources suggest that the pronunciation link in second link is also common. For example, this link: http://www.guidetojapanese.org/learn/grammar/hiragana provides pronunciation for /hu/.

So which one is correct, or both?

Note: I'm a Vietnamese native speaker and in our country, almost anyone can spot the difference between the two pronunciations in the links listed above.

Answer 1

It turns out that OP's question is that Japanese ふ sometimes sounds as if Vietnamese ph //f// and other times kh //x//.

That observation is true. The status of [[ɸ]] sound in Japanese is somewhat shaky because it appears mostly as an allophone in the environment //h// + //u// (strictly speaking, however, influx of modern loanwords has developed independent ファ行 to a quite degree).

[[ɸ]] in the Japanese language is a historical remnant of the times when all ハ行 were //ɸ//, which now only retained because //u// has labial feature. Hence, to some speakers the labiality is essential in pronunciation, while to others it is merely incidentally labialized //h//. In other words, the strength of labial narrowing varies among speakers.

Another factor is the realization of //h//. It is pronounced not only as [[h]] but also as [[x]] or [[χ]], especially in Eastern Japan including Tokyo. (Note that these allophones are natural in most regions when after consonantal element, cf. ゴッホ.) Since we don't have any contrastive fricatives in the back side of cavity, //h// can be freely realized from [[h]] (≈ Vietnamese h) to [[x]] (≈ Vietnamese kh).

Then we can draw a diagram to illustrate the gradation of ふ's consonant:

strong labial                      [ɸ]
     ↕
 weak labial           [hʷ]                    [xʷ]
                      glottal <--------------> velar

For example, in the sound clips of 船（ふね） on forvo.com, strawberrybrown's is closest to [[ɸ]], akitomo's closest to [[hʷ]], and kaoring's closest to [[xʷ]] (I couldn't get each direct link).

I'm not sure what is the distinctive features of ph in Vietnamese, but your first link of 冬 from jisho.org sounds rather an example of [[hʷ]] (though it doesn't seem to be a human articulation).

PS As an aside, it reminds me of a female announcement voice for some JR lines which utters ふつう with such strikingly dominant velar feature that almost sounds like くつう to me.

Answer 2

I found answer for it after days of searching on the internet.

Important note: The actual sound may vary depending on the accent of the speaker as you can find in the links that I posted in the questions.

Some people pronounce [ɸ] as in ふ the following way:

• Try to make a voiceless sound as if you are blowing a balloon (the [ɸ] appears here).
• Add ɯ sound (for ふ in this case)

The Youtube video links for the above pronunciation can be found here and here also.

Some other people pronounce the [ɸ] sound the same with [x] (or my preference is to use [kʰ] in this case instead), ref: https://en.wikipedia.org/wiki/Voiceless_velar_fricative.