Every student I teach practises this, even though almost everyone resists it at first because it feels silly. You stand in front of a mirror, you put your hands on your own face, and you let your hands trace the shape your mouth wants to make. Not the shape you think a vowel should look like. The shape it actually makes when nobody is managing it.
You start with the easiest sound there is, something so simple your body does not have to think about it: a meow. Just make the sound and pay attention to your mouth doing it, no analysis required. Then you let that unspool into a vowel run, moving through the open, resonant vowels in one smooth, connected line, each one flowing into the next with no gaps and no reset in between. The whole time your fingers rest on your lips and cheeks, following wherever your mouth goes instead of you deciding where it goes.
That is the exercise. None of it is arbitrary.
I have students touch their own face because your hands are doing two jobs at once. First, they are giving you physical feedback, an actual sensation of the shape happening, which is a completely different channel of information than trying to remember a mouth position from the inside. Touch and your own internal sense of where your body is run through different pathways than the kind of conscious verbal instruction you give yourself when you think “round the lips now.” You are handing the job to a system that reports back in sensation, not in words. Second, your hands are leading. That is the part people do not expect. You are not touching your face to check what your mouth already did. You are letting your hand start the movement and your mouth follow, because your hand does not have an opinion about what a vowel is supposed to look like. It just moves.
Pair that with watching yourself in the mirror and you have built a loop: hand leads, eyes confirm, mouth follows, and your brain never gets pulled into the driver’s seat to overthink the shape. The watching part is not vanity, by the way. There is a system in the brain that fires both when you perform a movement and when you see a movement happen, whether that movement belongs to you or to somebody across the room. It is a large part of how humans pick up physical skills by watching in the first place, turning something seen into something the body knows how to do. When you watch your own hand and mouth in the mirror while you are making the sound, you are feeding that same visual channel back into yourself in real time, reinforcing the shape a second way instead of relying on touch alone. That is the whole point. Your instrument already knows how to move. Your job is to stay out of its way long enough to let it.
This is also why I insist you do it with real feeling in it, not a flat, clinical run through the vowels like you are reading them off a chart. Go through it actually happy, eyebrows up, real smile, your whole face lit up, then do it again actually heavy or sad, and you will feel the shapes change even though it is technically the same handful of sounds. That is not decoration. There is a real, physical difference between a genuine expression and a performed one, and it shows up in exactly which muscles switch on. A polite, on command smile mostly uses the muscle at the corners of your mouth. A real one, the kind you cannot quite produce just by deciding to, pulls in the muscles around your eyes as well, the ones that crinkle the outer corners, and those simply do not fire when you are going through the motions for form’s sake. Your face is not one muscle with a single on off switch. It is a much wider, more connected network, and genuine emotion is the thing that recruits the whole network instead of the small, polite corner of it. Skip that and do the exercise flat, and you are training a narrower, more mechanical version of the movement, which is exactly the version that shows up on stage when nerves make everything tighten up anyway. Train it loose and expressive in the practice room and it has a much better chance of staying loose and expressive when it counts.
Then I have people run the sequence both directions, forward through the vowels and then backward, because that cross training matters more than people think. Each side of your brain has its own natural bias for handling a task like this, and running it both ways forces both hemispheres to get fluent instead of letting one quietly take over. The two sides of your brain stay in constant contact through a thick bundle of nerve fibers that does nothing but carry information back and forth between them, and that connection is what allows coordinated, whole body movement to happen at all. It is the same kind of wiring that lets your two hands work together on a task neither one could manage alone, sending a copy of the motor plan across so both sides stay synced on timing instead of guessing at each other. Singing with your whole face and body engaged is a whole body movement whether you think of it that way or not. When you can do the sequence forward and backward with equal ease, that is your sign that you have built real coordination instead of a memorized track running on autopilot down one side.
It’s bigger than clean vowels. Your body’s muscle memory is deeper and more durable than anything you memorize in your head, and that is not a figure of speech, that is just how motor learning works. It lives in a different, sturdier system than the one that stores facts and instructions, tucked into older, more automatic parts of the brain rather than the part that has to consciously search and recall. That is why a physical skill tends to survive long after the head has let go of the mental instructions that built it in the first place, and why it holds up under nerves and rust in a way that memorized facts do not. So once you stop mentally managing every vowel shape and let your body take over, something else stops too. You stop unconsciously borrowing someone else’s mouth shape, someone else’s placement, someone else’s way of curling into a note, because you are no longer running your voice through a mental reference library of singers you admire. During genuine, spontaneous performance, the part of the brain that handles self monitoring and self censoring has to quiet down before real expression can come through, and a mind still busy comparing itself to a reference recording is a mind that has not made room for that yet. You are just letting your own instrument move the way it actually moves.
Nobody needs another version of a voice that already exists. What people respond to is hearing someone who sounds like themselves, physically and honestly, with no imitation sitting between the feeling and the sound. That is not a small technical fix. That is the whole reason I start here.
Further reading: Fake Smile or Genuine Smile? The Duchenne Smile, Paul Ekman Group. The Effect of Modeling Methods on Mirror Neuron Activity and Motor Skill Acquisition.