|Voice onset time|
The diacritic for aspiration in the International Phonetic Alphabet is a superscript "h", . Unaspirated consonants are not normally marked explicitly, but there is a diacritic for non-aspiration in the Extensions to the IPA, the superscript equal sign, .
Voiceless consonants are produced with the vocal cords open and voiced consonants are produced when the vocal folds are fractionally closed. Voiceless aspiration occurs when the vocal cords remain open after a consonant is released. An easy way to measure this is by noting the consonant's voice onset time, as the voicing of a following vowel cannot begin until the vocal cords close.
English voiceless stop consonants are aspirated for most native speakers when they are word-initial or begin a stressed syllable, as in pen, ten, Ken. They are unaspirated for almost all speakers when immediately following word-initial s, as in spun, stun, skunk. After s elsewhere in a word they are normally unaspirated as well, except when the cluster is heteromorphemic and the stop belongs to an unbound morpheme; compare dis[t]end vs. dis[tʰ]aste. Word-final voiceless stops optionally aspirate.
Aspirated consonants are not always followed by vowels or other voiced sounds; indeed, in Eastern Armenian, aspiration is contrastive even at the ends of words. For example compare: pillow, with difficult and high.
In many languages, such as the Chinese languages, Indo-Aryan languages (from Sanskrit), Dravidian languages (i.e. under the influence of Sanskrit. Tamil, the classical Dravidian tongue does not show aspiration at all), Icelandic, Korean, Thai, and Ancient Greek, etc. and etc. are different phonemes altogether.
Alemannic German dialects have unaspirated as well as aspirated ; the latter series are usually viewed as consonant clusters. In Danish and most southern varieties of German, the "lenis" consonants transcribed for historical reasons as are distinguished from their "fortis" counterparts mainly in their lack of aspiration.
Icelandic and Faroese have pre-aspirated ; some scholars interpret these as consonant clusters as well. Preaspirated stops also occur in some Sami languages; e.g. in Skolt Sami the unvoiced stop phonemes , , , are pronounced preaspirated (, ) when they occur in medial or final position.
There are degrees of aspiration. Armenian and Cantonese have aspiration that lasts about as long as English aspirated stops, as well as unaspirated stops like Spanish. Korean has lightly aspirated stops that fall between the Armenian and Cantonese unaspirated and aspirated stops, as well as strongly aspirated stops whose aspiration lasts longer than that of Armenian or Cantonese. (See voice onset time.) An old IPA symbol for light aspiration was (that is, like a rotated ejective symbol), but this is no longer commonly used. There is no specific symbol for strong aspiration, but can be iconically doubled for, say, Korean * vs. *. Note however that Korean is nearly universally transcribed as vs. , with the details of voice onset time given numerically.
Aspiration also varies with place of articulation. Spanish /p t k/, for example, have voice onset times (VOTs) of about 5, 10, and 30 milliseconds, whereas English /p t k/ have VOTs of about 60, 70, and 80 ms. Korean has been measured at 20, 25, and 50 ms for and 90, 95, and 125 for .
The word 'aspiration' and the aspiration symbol is sometimes used with voiced stops, such as . However, such "voiced aspiration", also known as breathy voice or murmur, is less ambiguously transcribed with dedicated diacritics, either or . (Some linguists restrict the subscript diacritic to sonorants, such as vowels and nasal consonants, which are murmured throughout their duration, and use the superscript for the murmured release of obstruents.) When murmur is included under the term aspiration, "voiceless aspiration" is called just that to avoid ambiguity.