What Lip Reading Can and Can't Do

Logosofia no Dia a Dia · Sep 13, 2026 · 23m

Lip reading looks effortless in films and impossible in practice — and the truth is somewhere in between. This episode digs into speechreading: the clinical term for the full multimodal skill of reading lips, faces, gesture, and context together. We explore why only about 30–40% of speech sounds are visually distinguishable, how visemes collapse "pat," "bat," and "mat" into a single mouth shape, and why coarticulation turns speech into a continuous blur rather than discrete beads on a string. The McGurk effect shows visual speech isn't a backup channel — it's wired into hearing itself. We…