A viseme is the visible shape your mouth makes for a sound. Say "pop," "bob," and "mom" in a mirror. Your lips do exactly the same thing all three times, even though the sounds differ. That shared lip shape is one viseme covering three phonemes. Understanding this mapping is the fastest way to connect what pronunciation feels like inside your mouth with what it looks like from outside.
Quick answer
Visemes are mouth-shape units: each one groups the sounds that look identical on the lips (like /p, b, m/). English has around 40 phonemes but only about 12 to 15 visemes, which is why lip-reading alone is ambiguous, and why pairing visemes with audio feedback, the way Spelly does, trains pronunciation far faster than sound alone.
Watch your own mouth shapes
Record this sentence and compare Spelly's mouth visuals against your own lip shapes. Matching the shape is half of producing the sound.
The main English visemes, grouped
| Mouth shape | Phonemes sharing it | Example words |
|---|---|---|
| Lips pressed together | /p, b, m/ | pop, bob, mom |
| Teeth on lower lip | /f, v/ | fan, van |
| Tongue between teeth | /θ, ð/ | think, this |
| Tongue behind teeth ridge | /t, d, s, z, n, l/ | ten, den, sun, zoo, no, low |
| Lips rounded and pushed out | /w, uː, ɔː/ and sh-like /ʃ, tʃ/ | we, food, door, ship, chair |
| Jaw open, tongue low | /æ, ʌ, ɑː/ | cat, cup, car |
Why visemes matter more than most learners think
Here is the uncomfortable truth about pronunciation practice: you cannot see your own tongue. When a teacher says "round your lips more," you guess. When an app shows a mouth visual synced to your recording, you stop guessing. That visual channel carries information audio alone cannot, especially for sounds your native language does not have, where your ear literally cannot hear the difference yet.
This is also why Spelly shows both mouth and tongue visuals instead of mouth alone. The lips tell you /p/ from /f/, but only the tongue position separates /s/ from /ʃ/ or /l/ from /r/. Learners who train with both channels fix stubborn sounds weeks faster than learners training by ear only. Speech therapists have known this for decades. Language apps are only now catching up.
Three ways to train with visemes this week
- Mirror drill: pick one viseme group and say all its words watching your lips stay identical.
- Mute drill: watch a native clip on mute and guess the viseme sequence before unmuting.
- Mismatch hunt: record in Spelly and find every place your viseme differed from the reference.
Go deeper into the physical side with why articulation beats accent worry, the science in unlocking your vocal tract, and the sound-side companion IPA for pronunciation beginners.
FAQ
What is a viseme in simple words?
A viseme is the visible mouth shape for a speech sound. Sounds that look identical on the lips, like /p/, /b/, and /m/, share one viseme. English has about 40 sounds but only 12 to 15 visemes.
What is the difference between phonemes and visemes?
A phoneme is a unit of sound you hear. A viseme is a unit of mouth shape you see. The mapping is many-to-one: several phonemes can share one viseme, which is why lip-reading alone cannot identify every word.
Can visemes really help my pronunciation?
Yes, especially for sounds missing from your native language. When your ear cannot yet hear the difference, your eyes can see it. Pairing mouth visuals with audio feedback trains the sound from two directions at once.
See every sound you make
Spelly pairs every recording with mouth and tongue visuals synced to your voice. If you want to stop guessing what your mouth is doing, look at it while you speak.
Try Spelly for free