This question already has an answer here:
Ok, so I mostly understand how vowel devoicing works in Japanese, but there's something that's making me a bit confused. On all the sources I've found, most say that vowel devoicing usually occurs in a vowel sound in-between two voiceless consonants - specifically k, s, sh, t, ch, h, f and on rare occasions, apparently b, and p. However, there are a few times where I've seen a voiceless vowel occur with "d" as the second consonant, and at one point, I believe I've heard an "r" as the second consonant as well ( Examples: しだ、知らず ). Now, is it that maybe these are unique cases, or that I'm just not hearing them correctly? Or is there some other rule I'm missing?