How to Film a Talking Head on a Phone: 2026 Setup

Three changes do almost all of the work. Film on the 2x lens instead of the default wide, get a microphone within a foot of the speaker, and lock exposure and focus before you press record. Everything below is the detail behind those three, using the manufacturers' own published numbers.

None of it is about resolution. Your phone has out-resolved the delivery format for years. What still reads as phone video is the lens, the microphone distance, and the camera quietly re-metering halfway through your sentence.

Why does phone video look like phone video?

Three tells, in the order viewers notice them.

• The wide lens. A phone's default camera is a wide angle. At conversational distance it pushes your nose forward and pulls your ears back. People read the result as amateur without being able to name why.
• Room audio. The phone's own microphones are a metre away from your mouth, which means they capture the room at roughly the same level as the voice. Reverb is the single loudest amateur signal in the file.
• Auto exposure hunting. You lean forward, the frame brightens, the camera compensates, and the shot visibly breathes. It happens most on the sentence you most want to keep.

Which lens should you actually film on?

Phones have several rear cameras and the camera app hides which one you are using behind a zoom number. The figures below are Apple's published tech specs for the iPhone 17 Pro, introduced in 2025. Other models differ, so check your own spec page, but the shape of the answer is the same across recent flagships.

• 1x, Fusion Main: 24 mm, f/1.78, 48MP. The default and the wrong choice for a face at desk distance.
• 2x: 48 mm, f/1.78, 12MP at what Apple calls optical quality, taken from the middle of the main sensor. A standard portrait focal length on a lens that is just as fast as the main.
• 4x, Fusion Telephoto: 100 mm, f/2.8, 48MP. Flattering, but it needs a long room and it is about one and a third stops slower than the 2x, which you pay for in a dim office.
• 0.5x, Fusion Ultra Wide: 13 mm, f/2.2, 120 degree field of view. Never for a face.
• Front camera: 18MP, f/1.9, 4K Dolby Vision up to 60 fps. Better than it used to be and still behind every rear option.

Film on 2x. Back the phone up to about a metre and a half for a chest-up landscape frame, or nearer a metre if you are shooting vertical, since the taller frame fits the same subject from closer. Because the 2x on this generation comes out of the middle of the main sensor at the same f/1.78 aperture, you are not paying for the focal length in light, which is the usual reason people talk themselves back onto the wide lens.

What should you lock before you press record?

Five settings, once, and then never again.

• Resolution and frame rate: 4K at 30 fps. Apple lists 4K Dolby Vision at 24, 25, 30, 60, 100 and 120 fps on the 17 Pro. The high frame rates exist for slow motion. A person speaking at 120 fps is just a larger file.
• HDR video: turn it off in the camera format settings if the file is going anywhere other than an Apple device. Dolby Vision looks correct on the phone that shot it and unpredictable in a browser or a social feed, where it commonly renders washed out or crushed.
• Exposure and focus lock: press and hold on your face in the preview until the lock indicator appears. This is the setting that stops the breathing.
• Clean the lens. The cheapest fix on the list and the one skipped most often. A fingerprint on the glass turns every light source into a haze.
• Use the rear camera. If you need to see yourself, use a mirror behind the phone or a second screen, not the front camera.

Shoot 4K even though you will deliver 1080. The extra pixels let an editor punch in for a second angle out of a single take, which is the cheapest way to make one camera behave like two.

What does the microphone actually change?

More than the camera does. Apple lists four studio-quality mics and wind noise reduction on the 17 Pro, and they are genuinely good microphones. They are also a metre from your mouth, and distance is what makes audio sound like a room instead of a person.

A clip-on wireless transmitter moves the capsule to about a hand's width below your chin. Here is what one of the common ones publishes about itself, from DJI's own spec page for the Mic Mini.

• Transmitter: approximately 10 g excluding the clip magnet, 114 mAh battery, approximately 11.5 hours of operating time.
• Receiver: approximately 17.8 g, 170 mAh, approximately 10.5 hours.
• Charging case: 1950 mAh, and DJI states it will fully charge two transmitters and one receiver approximately 3.6 times.
• Polar pattern: omnidirectional. Frequency response 20 Hz to 20 kHz with the low cut off, or 100 Hz to 20 kHz with it on.
• Maximum sound pressure level 120 dB SPL, equivalent noise 24 dBA.
• Maximum transmission distance 400 m, which DJI qualifies as measured in an open, unobstructed, interference-free environment and for reference only.

Two of those numbers matter for a seated founder video and the rest are marketing. The omnidirectional pattern means the mic does not need to be aimed, so clipping it slightly off centre costs you nothing, and distance is the only variable you control. The low cut setting rolls off everything below 100 Hz, which removes desk rumble and air handling and takes nothing useful out of a speaking voice. Leave it on indoors.

Price has moved recently in the buyer's favour. Rode cut the Wireless Micro from $149 to $99, announced in September 2025 alongside reductions across its wireless range. At that price the microphone is the highest return purchase in this entire setup and it is not close.

Where does the light go?

One source, in front of you, slightly above eye line and off to one side. That is the whole recipe.

• A window is free and better than most panels people buy. Face it. Do not sit with it behind you, which turns you into a silhouette the camera then tries to correct.
• Turn the overheads off rather than trying to balance them against the window. Mixed colour temperature is harder to fix than underexposure.
• Check the nose shadow. If you can see a defined shadow crossing your lip, move the light closer to the camera axis until it disappears.
• Nothing bright behind you in frame. A lamp or a window in the background will pull the auto exposure even after you have locked it, because reflections change as you move.

How do you frame it?

Eyes on the upper third, camera at eye height, chest up. Eye height means the phone is on a tripod, not resting against a stack of books, because you will want to nudge it by two centimetres and a book stack does not do that.

If the clip is being cut vertical afterwards, check the framing against the platform overlays before you shoot rather than after. The published safe zones for Reels, TikTok and Shorts eat more of the frame than people expect, and a well framed chest-up shot can still lose its lower third to captions and buttons. The same applies to the export settings each platform actually publishes, which are worth setting once in your editor as a preset.

What does this cost end to end?

Less than the equipment conversation implies.

• The phone: you already own it.
• The light: free, if you sit facing a window during daylight.
• The microphone: $99 for a Rode Wireless Micro at the price announced in September 2025.
• A phone tripod with a clamp: the only other purchase, and the cheapest item in the setup.

That is the entire kit. Anything past it is buying comfort, not quality, until you are shooting often enough that comfort becomes speed.

What do you not need?

Four things that are real, impressive, and irrelevant to a founder talking to a camera for 90 seconds.

• ProRes RAW, Apple Log 2 and 4K at 120 fps. Apple lists all three on the 17 Pro, and notes that ProRes at 4K 120 fps requires external recording. Log footage also has to be graded before it looks like anything. This is real capability aimed at people finishing in a colour suite.
• A gimbal, for a seated shot. It solves movement you are not doing.
• A second camera, unless you own a second phone. Shooting 4K and punching in covers the same need.
• A teleprompter, for anything under 90 seconds. Reading is visible in the eyes, and the fix for forgetting your point is bullet points on a sticky note beside the lens.

If you want the video to have any cutaways at all, the shots that make that possible are worth planning before the shoot. A shot list matched to the beats of the script takes about two hours to capture once the interview is done.

Frequently asked questions

Is a phone good enough for a professional talking head video?

Yes, and it has been for several generations. Apple publishes 4K Dolby Vision capture with sensor-shift stabilisation on the 17 Pro. What separates amateur from professional output at this point is lens choice, microphone distance and lighting, none of which are camera limitations.

Should I use the front camera or the back camera?

The back camera, always. On the iPhone 17 Pro the front camera is an 18MP f/1.9 fixed wide, while the rear system gives you a 48 mm portrait focal length at f/1.78. Use a mirror or a second screen behind the phone if you need to monitor framing.

Do I need a wireless mic in a quiet room?

Quiet is not the issue, distance is. A silent room still has reverb, and the phone's microphones sit about a metre away, so they capture the reflections at close to the level of your voice. Moving the capsule under your chin fixes what silence cannot.

Should I shoot in 4K or 1080p?

Shoot 4K, deliver 1080. The surplus resolution lets an editor crop into the frame for a tighter second angle, so a single take yields two apparent camera positions. Storage is the only cost, and it is a smaller cost than reshooting.

Does shooting in Log make the video look better?

Not by itself. Log is a flat capture profile that preserves range for grading later. Ungraded, it looks washed out. It pays off when someone is colouring the footage properly, and costs you time and file size when nobody is.

Key takeaways

• Film on the 2x lens, not the default wide. On the iPhone 17 Pro that is a 48 mm equivalent at f/1.78, the same aperture as the main camera.
• Microphone distance beats microphone price. A clip-on transmitter under the chin removes the reverb that four studio-quality phone mics a metre away cannot.
• Lock exposure and focus before recording. Auto metering breathing is the most visible amateur tell after the lens choice.
• Turn HDR video off unless the file stays inside Apple's ecosystem, because Dolby Vision renders unpredictably in browsers and social feeds.
• One light source in front, slightly above and to one side. A window beats most panels and costs nothing.
• Shoot 4K and deliver 1080, so an editor can crop in and give one take a second angle.
• Skip Log, ProRes RAW, gimbals and teleprompters for a seated 90 second video. They are real features aimed at a different job.

When the filming stops being the hard part

The setup above takes an afternoon to get right and then repeats forever. What does not repeat cheaply is the edit, and that is usually where founder video actually stalls: a folder of perfectly good takes and nobody to cut them into something publishable this week. If you have reached that point, the published prices and options for founder video editing are the next thing worth reading.

EchoPulse runs that pipeline for founders and operators, taking raw footage through to published clips with senior review on every file and a 14 day paid pilot to start.