Visible speech is recognizable when the mouth is depicted by just a few pixels

What is it about?

Capitalizing on the McGurk effect (incongruent visemes of a speaker produce auditory illusory perception) and spatial quantization of the video image of the speaker it is demonstrated that for producing the illusioon it is sufficient to show the speaker head at a coarse level of pixelation.

Featured Image

Why is it important?

The results prove that audio-visual interaction in perceiving visible speech can use also quite coarse visual information and not neccessarily finne visual detail of the whole face and mouth.