In Tokyo dialect (standard Japanese), the sounds /i/ and /u/ usually become voiceless (i.e., whispered or devoiced) if they appear between voiceless consonants (/k/, /s/, /sh/, /t/, /ts/, /p/, or /h/) or at the end of a word.

If the sentence ends with /...desu/ (copula ending) or /...masu/ (polite verb ending), the final /u/ sound is also devoiced. If the sentence does not end with these two words, devoicing of the sentence-final /i/ and /u/ sounds may or may not occur. Compare the following:

	Literal pronunciation			Natural pronunciation
"history"	re ki shi れ き し			rekishi (or rekishi)	(or )
"munch"	pa ku tsu ku ぱ く つ く			pakutsuku (or pakutsuku)	(or )
"junk"	ga ra ku ta が ら く た			garakuta
"I am Brown."	bu ra u n desu ブラウンです			buraun desu

Devoicing phenomena are much more complex than what's described above, and there are many regional, phonological and situational factors to consider, but we will not go into those here. Listen to the native speakers carefully and learn by examples.

Words that look identical (when written in ひらがな or カタカナ) may have different meanings depending on the intonation. Since ひらがな, カタカナ as well as roomaji do not differentiate intonation contours, the only way to learn the differences is by listening carefully to the pronunciation of native speakers. In the following, the first きた is pronounced as the verb "came" and the second きた is pronounced as the verb "put on [clothes]". When syllables turn voiceless (き part), the differences become rather subtle.

"came" vs. "put on [clothes]"

vs.