Lip Reading

A recent study looked at the visual characteristics of human lip movements while we speak, in order to better understand common lip-reading errors.

Much of the existing research in this space focuses on phonemes, which are the sounds that make up words in a particular language. Lip readers commonly mix up phonemes, which can then lead to incorrect understandings of what’s being said.

This study instead focused on what the researchers behind it call visemes, which are the visual equivalent of phonemes: the movement and shape of a person’s lips, jaw, and mouth while they speak, rather than the sounds emitted.

While some words sound and look similar, like ‘kit,’ ‘cat,’ and ‘cut,’ others look similar (the lips, jaw, and mouth making similar movements) but sound different, like ‘vet,’ ‘fit,’ and ‘fuzz.’