You recorded a clip, ran it through auto captions, and half the corrections you had to make were words you said perfectly clearly. Or so it sounded to you. Most of those misses were decided before any caption app got involved, at the moment you pressed record: how far the phone was from your mouth, what the room did to your voice, and what else the microphone picked up.
The short answer. Get the iPhone’s microphone as close to your mouth as you can, record in a small room with soft furnishings, keep wind and fingers away from the mic holes, and silence notifications. If the phone has to be more than an arm’s length away, use a clip-on or wireless mic. Then listen back on headphones before you caption, because that’s closer to what the transcription hears.
Why the recording decides caption accuracy
Speech recognition works from the sound alone. When your voice is loud and clear relative to everything else on the track, it gets nearly every word. When the voice is competing with echo, traffic or a fridge hum, it starts guessing, and the guesses are plausible real words rather than obvious nonsense, which is why they’re easy to miss on a quick read.
The causes behind those guesses, from music to crosstalk, are covered in why auto captions get words wrong. This post is the practical half: what to do with an iPhone in your hand to stop them.
Distance is the biggest lever
Sound from your mouth spreads out as it travels, so its level drops quickly with distance. In open space, doubling the distance cuts the sound energy reaching the mic to about a quarter. The room noise and echo around you don’t drop the same way, because they’re coming from everywhere. So every step the phone moves away, your voice shrinks and the room stays put.
In practice that means:
- Selfie at arm’s length is fine in a quiet room. It’s already near the limit.
- Phone on a tripod across the room is where captions start to fall apart, even if the picture looks great.
- Phone propped just out of frame below your face often sounds better than the same phone filming you, because it’s closer.
If you want a wide shot and clean audio, the answer is a second microphone near your mouth, not a louder voice. Raising your voice changes how you sound on camera and only partly helps.
Know where the mics are
Recent iPhones have several microphones, usually including one on the bottom edge, one near the earpiece at the top and one beside the rear cameras. Which ones the Camera app uses can depend on the model and which camera you’re filming with. You don’t need the exact map. Just don’t cover any of the small holes with a finger, a thick case or a phone grip, and check that a new case lines up with them.
Pick the room before the shot
Hard, flat surfaces bounce your voice back to the mic a few milliseconds late. That echo smears one word into the next, and it’s one of the most common reasons clean-sounding speech still transcribes badly.
A quick way to check a room: clap once and listen. A sharp ring or a flutter means lots of reflections. A dull, short thud means the room is soaking up sound.
Rooms that usually work well:
- A bedroom with a bed, curtains and a rug
- A living room with a sofa and bookshelves
- Inside a parked car with the engine off
Rooms that usually don’t: bathrooms, kitchens, empty rooms after a move, stairwells and anywhere with a tiled floor. If you’re stuck in one, get closer to the mic and hang a duvet or a couple of coats behind the phone. It looks silly and it works.
Turn off what hums: fans, air conditioning, a dishwasher, the fridge if it’s in the same room. You stop noticing a steady hum after a minute. The microphone doesn’t.
Outdoors: wind, traffic and handling noise
Wind is the outdoor problem. It hits the mic directly and produces low, rumbling bursts that can drown whole words. Turn so the wind is behind the phone rather than blowing across it, use a wall or a car as a shield, and if you film outside often, a small clip-on mic with a fluffy windscreen fixes it far better than anything in software.
At the time of writing, some recent iPhone models also have a Wind Noise Reduction switch under Settings, then Camera, then Record Sound. It’s worth checking it’s on if your model has it, but it reduces wind rather than removing it.
Handling noise is the other one. Fingers moving on the phone, a ring tapping the case, and the click of switching grip all land right on the microphone. Hold the phone still while you talk, or put it on something.
Traffic and crowd noise are harder. Get close to the mic, keep your back to the noise if you can, and pause when a bus goes past. Re-recording one line costs less time than correcting it.
When an external mic is worth it
Two kinds are common for iPhone video:
| Type | Good for | Watch out for |
|---|---|---|
| Wired clip-on (lavalier) | Talking to camera, tutorials, one speaker | Cable in shot, needs the right connector or adapter |
| Wireless kit (clip-on transmitter, receiver in the phone port) | Wide shots, walking, two people with two transmitters | Batteries, and checking the receiver is actually in use |
Clip the mic on your chest about a hand’s width below your chin, and keep clothing from rubbing against it. With any external mic, record a short test and play it back. If it sounds like the phone’s own mic, the phone may not be using it.
Settings and habits that help
- Silence the phone. Turn on a Focus or Do Not Disturb so a notification sound or vibration doesn’t land mid-sentence.
- Check Audio Mix after recording. At the time of writing, iPhone 16 models and later offer an Audio Mix option when you edit a video in Photos, with modes that can push background sound down. Try it before captioning a noisy clip, and listen to the result.
- Leave a second of silence at each end. It gives you room to trim and keeps the first word from getting clipped.
- Say tricky words cleanly. Names, brands and numbers are where models slip most, so don’t rush them.
A ten-second test before every shoot
- Record ten seconds of yourself talking, in the spot and framing you plan to use.
- Play it back on headphones, not the phone speaker.
- Listen for echo, hum, wind and your voice level.
- Fix the worst one and record again.
Once the audio is clean, captioning is the easy part. In Capster you import the clip and only the audio track is sent for transcription, so a clean recording is directly what decides the result. The guide to adding captions to a video on iPhone walks through the rest, and if you’re still choosing a tool, the auto subtitle app for iPhone page covers what to look for.