Voicing is not phonemic in either stops or fricatives; in particular, /β ð ɣ/ are always voiced, while /p t t͡s s t͡ɕ ɕ k/ are always voiceless.

Myrkçam may seem somwhat atypical in that it observes no voicing distinction (a feature shared by about a third of all languages); nonetheless, both voiceless and voiced consonants occur, in complementary distribution. /m/, /n/, /ɲ/, /ŋ/, /j/, /w/, and /r/ are always voiced; all other consonants are voiceless except when they fall between two other voiced phonemes or when they fall immediately after a nasal plosive, in which case they become voiced. This rule is not applied recursively.

Though, plosives and fricatives voiced counterparts occur as allophones in complementary distribution in intervocalic position.

Mid-word consonants clusters are voiced too unless they be comprised of a /l/ and a plosive/fricative (in any order), then in this case the allophonic voicing phenomenon described above isn't mandatory but subject to free variation.

An exception is also the mid-word consonants cluster /n/ + /t/, where /t/ might not be voiced either.