Whispering is an unvoiced mode of phonation in which the vocal cords are abducted so that they do not vibrate; air passes between the arytenoid cartilages to create audible turbulence during speech. Supralaryngeal articulation remains the same as in normal speech.

In normal speech, the vocal cords alternate between states of voice and voicelessness. In whispering, only the voicing segments change, so that the vocal cords alternate between whisper and voicelessness (though the acoustic difference between the two states is minimal). Because of this, implementing speech recognition for whispered speech is more difficult, as the characteristic spectral range needed to detect syllables and words is not given through the total absence of tone. More advanced techniques such as neural networks may be used, however, as is done by Amazon Alexa.

There is no symbol in the IPA for whispered phonation, since it is not used phonemically in any language. However, a sub-dot under phonemically voiced segments is sometimes seen in the literature, as [ʃʊ̣ḍ] for whispered should.