London:

The utterance is produced with a pitch high in the speaker’s range. It is quiet and fast, until it gets to the word ‘cook’. On the word ‘cook’, there is a fall in pitch, and then the pitch remains low to the end of the question. In the fast talk that precedes the stressed word ‘cook’, there are a number of unusual sounds. The first one is [s9]. The diacritic [9] marks that the articulation is closer than close approximation: that is to say, the tongue tip is raised quickly, so that friction starts abruptly. This articulation is similar to that of the Irish ‘slit-t’ we looked at earlier, but has a sharper, more [s]-like quality; you could think of it as a ‘fast [t]’, which in some sense it is.

The sound transcribed  at the start of the word ‘gonna’ is made with voicing + friction + velarity. It is achieved in the same way that  is: that is, the tongue body is raised towards the velum, as if to make a velar closure,  , but it fails to make a complete closure and instead generates a short period of friction,  . As with  , the speed with which the articulation is made and its short duration both contribute to the perception of these sounds as [t] and   respectively.

Here is another example from English spoken in the north west of England (around Manchester), although this pattern is common in other varieties of English too (Kelly 1967). In clusters of the shape [-sPs], where P stands for a voiceless plosive [p t k], complete closure is not made; what happens is that two articulators approximate each other, producing a short period of friction. In the case of labials and dorsals, the friction is concurrent with alveolar friction. In the case of coronals, there is a slight change in the quality of the friction as the tongue is raised: