What you hear depends on what you see

A voice saying “ba”, dubbed onto a face mouthing “ga”, is often heard as “da”. The syllable is a compromise the brain invented.

McGurk and MacDonald published the mismatch in 1976. Vision of the lips and audition of the voice are fused into one phonetic object. When they disagree, the fusion is often a third consonant that neither source produced. Close your eyes and the soundtrack returns to ba.

It works on people who know the trick. Speech is audiovisual by default. The telephone is the impoverished case, not the natural one.