Compare the stops in the following pairs of English words:
The initial stops of the first set (pill, till, kill) are followed by a short puff of air which is not present after the stops in the second set of words. (It might be easier to notice the difference between the two sets of words if you speak them with your lips a short distance away from a candle, watching which words cause the flame to flicker more. Or else hold a piece of paper loosely in front of your mouth and watch which words cause the paper to flutter.)
This brief puff of air is called aspiration, and plosives which are followed by it are said to be aspirated. The IPA diacritic for aspiration is a superscript [h]:
The presence or absence of aspiration will not change the meaning of English words. [spʰɪl] still means 'spill', though it is a decidedly odd pronunciation. (Saying [pɪl] instead of [pʰɪl] increases the odds that your listener will mistakenly hear you as saying Bill.)
Where an English speaker does and doesn't use aspiration is predictable. For most English dialects, the two environments where voiceless plosives are aspirated are:
So aspiration can be used as another way of telling if some syllables are stressed or not. We can tell that the final -to of potato is completely unstressed (even though it has an [o] instead of a schwa), because its t has becomed tapped (as we expect in an unstressed syllable) instead of aspirated (as we expect in a stressed syllable). On the otherhand, the final syllable does have secondary stress in undertow, and its t is aspirated: [ˈʌndɹ̩ ˌtʰo].
Plosives are not aspirated in French. The English word two and the French word tout 'all' might both be given the broad transcription [tu], but they differ in the presence or absence of aspiration:
One of the most characteristic features of an English accent in French is aspirating plosives which should not be aspirated. Similarly, failure to aspirate plosives in the appropriate environments can contribute to a French accent in speaking English. In many languages, aspiration can change the meaning of a word. For these languages, the h-diacritic would need to be included even in broad transcriptions.
When a voiceless unaspirated plosive is followed by a vowel, the time when the vocal folds begin vibrating for the vowel will coincide almost exactly with the time when the plosive is released (give or take up to 20 milliseconds).
After a voiceless aspirated stop, however, the vocal folds will not begin vibrating until well after the plosive is released. There is a period of time when the vocal tract is producing neither the plosive nor the following vowel -- this is the puff of air.
These are just two of a number of different possible ways of coordinating the timing between vocal fold vibration and a closure in the mouth. Various languages make use of many points along this Voice Onset Time (or VOT) continuum.
Languages that make voicing contrasts usually choose two (or often three) points along this continuum. English has chosen to use position 2 for its voiced sounds and either 3 or 4 (depending on position in the word or syllable) for voiceless sounds. French has chosen to use 1 (fully voiced) and 3 (voiceless unaspirated). Mandarin Chinese has chosen to use 3 (voiceless unaspirated) and 5 (strongly aspirated).
If an aspirated plosive is followed an approximant, as in pray, the period of voicelessness after the plosive's release will coincide with the approximant and make it voiceless: